DSND-Arvato Project Workbook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Capstone Project: Create a Customer Segmentation Report for Arvato Financial Services\n",
    "\n",
    "In this project, you will analyze demographics data for customers of a mail-order sales company in Germany, comparing it against demographics information for the general population. You'll use unsupervised learning techniques to perform customer segmentation, identifying the parts of the population that best describe the core customer base of the company. Then, you'll apply what you've learned on a third dataset with demographics information for targets of a marketing campaign for the company, and use a model to predict which individuals are most likely to convert into becoming customers for the company. The data that you will use has been provided by our partners at Bertelsmann Arvato Analytics, and represents a real-life data science task.\n",
    "\n",
    "If you completed the first term of this program, you will be familiar with the first part of this project, from the unsupervised learning project. The versions of those two datasets used in this project will include many more features and has not been pre-cleaned. You are also free to choose whatever approach you'd like to analyzing the data rather than follow pre-determined steps. In your work on this project, make sure that you carefully document your steps and decisions, since your main deliverable for this project will be a blog post reporting your findings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries \n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import time\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import figure\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.decomposition import PCA\n",
    "# from sklearn.preprocessing import Imputer\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "# magic word for producing visualizations in notebook\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 0: Get to Know the Data\n",
    "\n",
    "There are four data files associated with this project:\n",
    "\n",
    "- `Udacity_AZDIAS_052018.csv`: Demographics data for the general population of Germany; 891 211 persons (rows) x 366 features (columns).\n",
    "- `Udacity_CUSTOMERS_052018.csv`: Demographics data for customers of a mail-order company; 191 652 persons (rows) x 369 features (columns).\n",
    "- `Udacity_MAILOUT_052018_TRAIN.csv`: Demographics data for individuals who were targets of a marketing campaign; 42 982 persons (rows) x 367 (columns).\n",
    "- `Udacity_MAILOUT_052018_TEST.csv`: Demographics data for individuals who were targets of a marketing campaign; 42 833 persons (rows) x 366 (columns).\n",
    "\n",
    "Each row of the demographics files represents a single person, but also includes information outside of individuals, including information about their household, building, and neighborhood. Use the information from the first two files to figure out how customers (\"CUSTOMERS\") are similar to or differ from the general population at large (\"AZDIAS\"), then use your analysis to make predictions on the other two files (\"MAILOUT\"), predicting which recipients are most likely to become a customer for the mail-order company.\n",
    "\n",
    "The \"CUSTOMERS\" file contains three extra columns ('CUSTOMER_GROUP', 'ONLINE_PURCHASE', and 'PRODUCT_GROUP'), which provide broad information about the customers depicted in the file. The original \"MAILOUT\" file included one additional column, \"RESPONSE\", which indicated whether or not each recipient became a customer of the company. For the \"TRAIN\" subset, this column has been retained, but in the \"TEST\" subset it has been removed; it is against that withheld column that your final predictions will be assessed in the Kaggle competition.\n",
    "\n",
    "Otherwise, all of the remaining columns are the same between the three data files. For more information about the columns depicted in the files, you can refer to two Excel spreadsheets provided in the workspace. [One of them](./DIAS Information Levels - Attributes 2017.xlsx) is a top-level list of attributes and descriptions, organized by informational category. [The other](./DIAS Attributes - Values 2017.xlsx) is a detailed mapping of data values for each feature in alphabetical order.\n",
    "\n",
    "In the below cell, we've provided some initial code to load in the first two datasets. Note for all of the `.csv` data files in this project that they're semicolon (`;`) delimited, so an additional argument in the [`read_csv()`](https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_csv.html) call has been included to read in the data properly. Also, considering the size of the datasets, it may take some time for them to load completely.\n",
    "\n",
    "You'll notice when the data is loaded in that a warning message will immediately pop up. Before you really start digging into the modeling and analysis, you're going to need to perform some cleaning. Take some time to browse the structure of the data and look over the informational spreadsheets to understand the data values. Make some decisions on which features to keep, which features to drop, and if any revisions need to be made on data formats. It'll be a good idea to create a function with pre-processing steps, since you'll need to clean all of the datasets before you work with them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  load in the data\n",
    "azdias = pd.read_csv('../../data/Term2/capstone/arvato_data/Udacity_AZDIAS_052018.csv', sep=';')\n",
    "customers = pd.read_csv('../../data/Term2/capstone/arvato_data/Udacity_CUSTOMERS_052018.csv', sep=';')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Memory Reduction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Due to the huge size of data, analysing the data is very time consuming. As such I started looking for solution online. I found a nice article \"[How to handle BigData Files on Low Memory?](https://towardsdatascience.com/how-to-learn-from-bigdata-files-on-low-memory-incremental-learning-d377282d38ff#e618)\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [],
   "source": [
    "def reduce_mem_usage(df):\n",
    "    \"\"\" \n",
    "    iterate through all the columns of a dataframe and \n",
    "    modify the data type to reduce memory usage.        \n",
    "    \"\"\"\n",
    "    start_mem = df.memory_usage().sum() / 1024**2\n",
    "    print(('Memory usage of dataframe is {:.2f}' \n",
    "                     'MB').format(start_mem))\n",
    "    \n",
    "    for col in df.columns:\n",
    "        col_type = df[col].dtype\n",
    "        \n",
    "        if col_type != object:\n",
    "            c_min = df[col].min()\n",
    "            c_max = df[col].max()\n",
    "            if str(col_type)[:3] == 'int':\n",
    "                if c_min > np.iinfo(np.int8).min and c_max <\\\n",
    "                  np.iinfo(np.int8).max:\n",
    "                    df[col] = df[col].astype(np.int8)\n",
    "                elif c_min > np.iinfo(np.int16).min and c_max <\\\n",
    "                   np.iinfo(np.int16).max:\n",
    "                    df[col] = df[col].astype(np.int16)\n",
    "                elif c_min > np.iinfo(np.int32).min and c_max <\\\n",
    "                   np.iinfo(np.int32).max:\n",
    "                    df[col] = df[col].astype(np.int32)\n",
    "                elif c_min > np.iinfo(np.int64).min and c_max <\\\n",
    "                   np.iinfo(np.int64).max:\n",
    "                    df[col] = df[col].astype(np.int64)  \n",
    "            else:\n",
    "                if c_min > np.finfo(np.float16).min and c_max <\\\n",
    "                   np.finfo(np.float16).max:\n",
    "                    df[col] = df[col].astype(np.float16)\n",
    "                elif c_min > np.finfo(np.float32).min and c_max <\\\n",
    "                   np.finfo(np.float32).max:\n",
    "                    df[col] = df[col].astype(np.float32)\n",
    "                else:\n",
    "                    df[col] = df[col].astype(np.float64)\n",
    "        else:\n",
    "            df[col] = df[col].astype('category')\n",
    "    end_mem = df.memory_usage().sum() / 1024**2\n",
    "    print(('Memory usage after optimization is: {:.2f}' \n",
    "                              'MB').format(end_mem))\n",
    "    print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) \n",
    "                                             / start_mem))\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reduce Memory of all dataset\n",
    "azdias_reduce = reduce_mem_usage(azdias)\n",
    "customers_reduce = reduce_mem_usage(customers)\n",
    "mailout_train_reduce = reduce_mem_usage(mailout_train)\n",
    "mailout_test_reduce = reduce_mem_usage(mailout_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the reduced Memory file\n",
    "azdias_reduce.to_csv('Udacity_AZDIAS_reduce_mem_052018.csv')\n",
    "customers_reduce.to_csv('Udacity_CUSTOMERS_reduce_mem_052018.csv')\n",
    "mailout_train_reduce.to_csv('Udacity_MAILOUT_052018_TRAIN_reduce_mem.csv')\n",
    "mailout_test_reduce.to_csv('Udacity_MAILOUT_052018_TEST_reduce_mem.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading memory reduced data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3071: DtypeWarning: Columns (19,20) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    }
   ],
   "source": [
    "# Read data with after memory reduction\n",
    "azdias = pd.read_csv('Udacity_AZDIAS_reduce_mem_052018.csv')\n",
    "customers = pd.read_csv('Udacity_CUSTOMERS_reduce_mem_052018.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read information on the attributes \n",
    "# top-level list of attributes and descriptions, organized by informational category\n",
    "attributes_des = pd.read_excel('DIAS Information Levels - Attributes 2017.xlsx', header=1)\n",
    "# Detailed mapping of data values for each feature in alphabetical order\n",
    "attributes_info = pd.read_excel('DIAS Attributes - Values 2017.xlsx', header=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VHN</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>910215</td>\n",
       "      <td>-1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>910220</td>\n",
       "      <td>-1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>21.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>910225</td>\n",
       "      <td>-1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>910226</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>910241</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 367 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0     LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  \\\n",
       "0           0  910215        -1         NaN       NaN          NaN   \n",
       "1           1  910220        -1         9.0       0.0          NaN   \n",
       "2           2  910225        -1         9.0      17.0          NaN   \n",
       "3           3  910226         2         1.0      13.0          NaN   \n",
       "4           4  910241        -1         1.0      20.0          NaN   \n",
       "\n",
       "   ALTER_KIND2  ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ...  VHN  \\\n",
       "0          NaN          NaN          NaN                   NaN  ...  NaN   \n",
       "1          NaN          NaN          NaN                  21.0  ...  4.0   \n",
       "2          NaN          NaN          NaN                  17.0  ...  2.0   \n",
       "3          NaN          NaN          NaN                  13.0  ...  0.0   \n",
       "4          NaN          NaN          NaN                  14.0  ...  2.0   \n",
       "\n",
       "   VK_DHT4A  VK_DISTANZ  VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008  WOHNLAGE  \\\n",
       "0       NaN         NaN      NaN             NaN             NaN       NaN   \n",
       "1       8.0        11.0     10.0             3.0             9.0       4.0   \n",
       "2       9.0         9.0      6.0             3.0             9.0       2.0   \n",
       "3       7.0        10.0     11.0             NaN             9.0       7.0   \n",
       "4       3.0         5.0      4.0             2.0             9.0       3.0   \n",
       "\n",
       "   ZABEOTYP ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0         3         1                    2  \n",
       "1         5         2                    1  \n",
       "2         5         2                    3  \n",
       "3         3         2                    4  \n",
       "4         4         1                    3  \n",
       "\n",
       "[5 rows x 367 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# View the data\n",
    "azdias.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VHN</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>891221.000000</td>\n",
       "      <td>8.912210e+05</td>\n",
       "      <td>891221.000000</td>\n",
       "      <td>817722.000000</td>\n",
       "      <td>817722.000000</td>\n",
       "      <td>81058.000000</td>\n",
       "      <td>29499.000000</td>\n",
       "      <td>6170.000000</td>\n",
       "      <td>1205.000000</td>\n",
       "      <td>628274.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>770025.000000</td>\n",
       "      <td>815304.000000</td>\n",
       "      <td>815304.000000</td>\n",
       "      <td>815304.000000</td>\n",
       "      <td>783619.000000</td>\n",
       "      <td>817722.000000</td>\n",
       "      <td>798073.000000</td>\n",
       "      <td>891221.000000</td>\n",
       "      <td>891221.000000</td>\n",
       "      <td>891221.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>445610.000000</td>\n",
       "      <td>6.372630e+05</td>\n",
       "      <td>-0.358435</td>\n",
       "      <td>4.421928</td>\n",
       "      <td>10.864126</td>\n",
       "      <td>11.745392</td>\n",
       "      <td>13.402658</td>\n",
       "      <td>14.476013</td>\n",
       "      <td>15.089627</td>\n",
       "      <td>13.700717</td>\n",
       "      <td>...</td>\n",
       "      <td>2.417322</td>\n",
       "      <td>6.001214</td>\n",
       "      <td>7.532130</td>\n",
       "      <td>5.945972</td>\n",
       "      <td>3.933406</td>\n",
       "      <td>7.908791</td>\n",
       "      <td>4.052836</td>\n",
       "      <td>3.362438</td>\n",
       "      <td>1.522098</td>\n",
       "      <td>2.777398</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>257273.486465</td>\n",
       "      <td>2.572735e+05</td>\n",
       "      <td>1.198724</td>\n",
       "      <td>3.638805</td>\n",
       "      <td>7.639683</td>\n",
       "      <td>4.097660</td>\n",
       "      <td>3.243300</td>\n",
       "      <td>2.712427</td>\n",
       "      <td>2.452932</td>\n",
       "      <td>5.079849</td>\n",
       "      <td>...</td>\n",
       "      <td>1.166572</td>\n",
       "      <td>2.856091</td>\n",
       "      <td>3.247789</td>\n",
       "      <td>2.771464</td>\n",
       "      <td>1.964701</td>\n",
       "      <td>1.923137</td>\n",
       "      <td>1.949539</td>\n",
       "      <td>1.352704</td>\n",
       "      <td>0.499512</td>\n",
       "      <td>1.068775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.916530e+05</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>222805.000000</td>\n",
       "      <td>4.144580e+05</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>445610.000000</td>\n",
       "      <td>6.372630e+05</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>668415.000000</td>\n",
       "      <td>8.600680e+05</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>891220.000000</td>\n",
       "      <td>1.082873e+06</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>21.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 361 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Unnamed: 0           LNR       AGER_TYP     AKT_DAT_KL  \\\n",
       "count  891221.000000  8.912210e+05  891221.000000  817722.000000   \n",
       "mean   445610.000000  6.372630e+05      -0.358435       4.421928   \n",
       "std    257273.486465  2.572735e+05       1.198724       3.638805   \n",
       "min         0.000000  1.916530e+05      -1.000000       1.000000   \n",
       "25%    222805.000000  4.144580e+05      -1.000000       1.000000   \n",
       "50%    445610.000000  6.372630e+05      -1.000000       3.000000   \n",
       "75%    668415.000000  8.600680e+05      -1.000000       9.000000   \n",
       "max    891220.000000  1.082873e+06       3.000000       9.000000   \n",
       "\n",
       "            ALTER_HH   ALTER_KIND1   ALTER_KIND2  ALTER_KIND3  ALTER_KIND4  \\\n",
       "count  817722.000000  81058.000000  29499.000000  6170.000000  1205.000000   \n",
       "mean       10.864126     11.745392     13.402658    14.476013    15.089627   \n",
       "std         7.639683      4.097660      3.243300     2.712427     2.452932   \n",
       "min         0.000000      2.000000      2.000000     4.000000     7.000000   \n",
       "25%         0.000000      8.000000     11.000000    13.000000    14.000000   \n",
       "50%        13.000000     12.000000     14.000000    15.000000    15.000000   \n",
       "75%        17.000000     15.000000     16.000000    17.000000    17.000000   \n",
       "max        21.000000     18.000000     18.000000    18.000000    18.000000   \n",
       "\n",
       "       ALTERSKATEGORIE_FEIN  ...            VHN       VK_DHT4A     VK_DISTANZ  \\\n",
       "count         628274.000000  ...  770025.000000  815304.000000  815304.000000   \n",
       "mean              13.700717  ...       2.417322       6.001214       7.532130   \n",
       "std                5.079849  ...       1.166572       2.856091       3.247789   \n",
       "min                0.000000  ...       0.000000       1.000000       1.000000   \n",
       "25%               11.000000  ...       2.000000       3.000000       5.000000   \n",
       "50%               14.000000  ...       2.000000       6.000000       8.000000   \n",
       "75%               17.000000  ...       3.000000       9.000000      10.000000   \n",
       "max               25.000000  ...       4.000000      11.000000      13.000000   \n",
       "\n",
       "             VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008       WOHNLAGE  \\\n",
       "count  815304.000000   783619.000000   817722.000000  798073.000000   \n",
       "mean        5.945972        3.933406        7.908791       4.052836   \n",
       "std         2.771464        1.964701        1.923137       1.949539   \n",
       "min         1.000000        0.000000        1.000000       0.000000   \n",
       "25%         4.000000        2.000000        8.000000       3.000000   \n",
       "50%         6.000000        4.000000        9.000000       3.000000   \n",
       "75%         8.000000        6.000000        9.000000       5.000000   \n",
       "max        11.000000        6.000000        9.000000       8.000000   \n",
       "\n",
       "            ZABEOTYP      ANREDE_KZ  ALTERSKATEGORIE_GROB  \n",
       "count  891221.000000  891221.000000         891221.000000  \n",
       "mean        3.362438       1.522098              2.777398  \n",
       "std         1.352704       0.499512              1.068775  \n",
       "min         1.000000       1.000000              1.000000  \n",
       "25%         3.000000       1.000000              2.000000  \n",
       "50%         3.000000       2.000000              3.000000  \n",
       "75%         4.000000       2.000000              4.000000  \n",
       "max         6.000000       2.000000              9.000000  \n",
       "\n",
       "[8 rows x 361 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>PRODUCT_GROUP</th>\n",
       "      <th>CUSTOMER_GROUP</th>\n",
       "      <th>ONLINE_PURCHASE</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>9626</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3</td>\n",
       "      <td>COSMETIC_AND_FOOD</td>\n",
       "      <td>MULTI_BUYER</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>9628</td>\n",
       "      <td>-1</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>FOOD</td>\n",
       "      <td>SINGLE_BUYER</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>143872</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>11.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>COSMETIC_AND_FOOD</td>\n",
       "      <td>MULTI_BUYER</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>143873</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "      <td>COSMETIC</td>\n",
       "      <td>MULTI_BUYER</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>143874</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1</td>\n",
       "      <td>FOOD</td>\n",
       "      <td>MULTI_BUYER</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 370 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0     LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  \\\n",
       "0           0    9626         2         1.0      10.0          NaN   \n",
       "1           1    9628        -1         9.0      11.0          NaN   \n",
       "2           2  143872        -1         1.0       6.0          NaN   \n",
       "3           3  143873         1         1.0       8.0          NaN   \n",
       "4           4  143874        -1         1.0      20.0          NaN   \n",
       "\n",
       "   ALTER_KIND2  ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ...  VK_ZG11  \\\n",
       "0          NaN          NaN          NaN                  10.0  ...      2.0   \n",
       "1          NaN          NaN          NaN                   NaN  ...      3.0   \n",
       "2          NaN          NaN          NaN                   0.0  ...     11.0   \n",
       "3          NaN          NaN          NaN                   8.0  ...      2.0   \n",
       "4          NaN          NaN          NaN                  14.0  ...      4.0   \n",
       "\n",
       "   W_KEIT_KIND_HH  WOHNDAUER_2008  WOHNLAGE  ZABEOTYP      PRODUCT_GROUP  \\\n",
       "0             6.0             9.0       7.0         3  COSMETIC_AND_FOOD   \n",
       "1             0.0             9.0       NaN         3               FOOD   \n",
       "2             6.0             9.0       2.0         3  COSMETIC_AND_FOOD   \n",
       "3             NaN             9.0       7.0         1           COSMETIC   \n",
       "4             2.0             9.0       3.0         1               FOOD   \n",
       "\n",
       "   CUSTOMER_GROUP  ONLINE_PURCHASE ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0     MULTI_BUYER                0         1                    4  \n",
       "1    SINGLE_BUYER                0         1                    4  \n",
       "2     MULTI_BUYER                0         2                    4  \n",
       "3     MULTI_BUYER                0         1                    4  \n",
       "4     MULTI_BUYER                0         1                    3  \n",
       "\n",
       "[5 rows x 370 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>ONLINE_PURCHASE</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>191652.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>145056.000000</td>\n",
       "      <td>145056.000000</td>\n",
       "      <td>11766.000000</td>\n",
       "      <td>5100.000000</td>\n",
       "      <td>1275.000000</td>\n",
       "      <td>236.000000</td>\n",
       "      <td>139810.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>143781.000000</td>\n",
       "      <td>143781.000000</td>\n",
       "      <td>143781.000000</td>\n",
       "      <td>137910.000000</td>\n",
       "      <td>145056.000000</td>\n",
       "      <td>141725.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>95825.500000</td>\n",
       "      <td>95826.500000</td>\n",
       "      <td>0.344359</td>\n",
       "      <td>1.747525</td>\n",
       "      <td>11.352009</td>\n",
       "      <td>12.337243</td>\n",
       "      <td>13.672353</td>\n",
       "      <td>14.647059</td>\n",
       "      <td>15.377119</td>\n",
       "      <td>10.331579</td>\n",
       "      <td>...</td>\n",
       "      <td>4.374417</td>\n",
       "      <td>4.564769</td>\n",
       "      <td>3.168868</td>\n",
       "      <td>4.152716</td>\n",
       "      <td>8.646371</td>\n",
       "      <td>3.723133</td>\n",
       "      <td>2.576806</td>\n",
       "      <td>0.090247</td>\n",
       "      <td>1.376432</td>\n",
       "      <td>3.060907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>55325.311233</td>\n",
       "      <td>55325.311233</td>\n",
       "      <td>1.391672</td>\n",
       "      <td>1.966334</td>\n",
       "      <td>6.275026</td>\n",
       "      <td>4.006050</td>\n",
       "      <td>3.243335</td>\n",
       "      <td>2.753787</td>\n",
       "      <td>2.307653</td>\n",
       "      <td>4.134828</td>\n",
       "      <td>...</td>\n",
       "      <td>2.924355</td>\n",
       "      <td>2.887035</td>\n",
       "      <td>2.233516</td>\n",
       "      <td>1.974375</td>\n",
       "      <td>1.154001</td>\n",
       "      <td>2.095540</td>\n",
       "      <td>1.168486</td>\n",
       "      <td>0.286536</td>\n",
       "      <td>0.484492</td>\n",
       "      <td>1.086254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>47912.750000</td>\n",
       "      <td>47913.750000</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>95825.500000</td>\n",
       "      <td>95826.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>143738.250000</td>\n",
       "      <td>143739.250000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>191651.000000</td>\n",
       "      <td>191652.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>21.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>18.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 362 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Unnamed: 0            LNR       AGER_TYP     AKT_DAT_KL  \\\n",
       "count  191652.000000  191652.000000  191652.000000  145056.000000   \n",
       "mean    95825.500000   95826.500000       0.344359       1.747525   \n",
       "std     55325.311233   55325.311233       1.391672       1.966334   \n",
       "min         0.000000       1.000000      -1.000000       1.000000   \n",
       "25%     47912.750000   47913.750000      -1.000000       1.000000   \n",
       "50%     95825.500000   95826.500000       0.000000       1.000000   \n",
       "75%    143738.250000  143739.250000       2.000000       1.000000   \n",
       "max    191651.000000  191652.000000       3.000000       9.000000   \n",
       "\n",
       "            ALTER_HH   ALTER_KIND1  ALTER_KIND2  ALTER_KIND3  ALTER_KIND4  \\\n",
       "count  145056.000000  11766.000000  5100.000000  1275.000000   236.000000   \n",
       "mean       11.352009     12.337243    13.672353    14.647059    15.377119   \n",
       "std         6.275026      4.006050     3.243335     2.753787     2.307653   \n",
       "min         0.000000      2.000000     2.000000     5.000000     8.000000   \n",
       "25%         8.000000      9.000000    11.000000    13.000000    14.000000   \n",
       "50%        11.000000     13.000000    14.000000    15.000000    16.000000   \n",
       "75%        16.000000     16.000000    16.000000    17.000000    17.000000   \n",
       "max        21.000000     18.000000    18.000000    18.000000    18.000000   \n",
       "\n",
       "       ALTERSKATEGORIE_FEIN  ...       VK_DHT4A     VK_DISTANZ        VK_ZG11  \\\n",
       "count         139810.000000  ...  143781.000000  143781.000000  143781.000000   \n",
       "mean              10.331579  ...       4.374417       4.564769       3.168868   \n",
       "std                4.134828  ...       2.924355       2.887035       2.233516   \n",
       "min                0.000000  ...       1.000000       1.000000       1.000000   \n",
       "25%                9.000000  ...       2.000000       2.000000       1.000000   \n",
       "50%               10.000000  ...       4.000000       4.000000       3.000000   \n",
       "75%               13.000000  ...       7.000000       7.000000       4.000000   \n",
       "max               25.000000  ...      11.000000      13.000000      11.000000   \n",
       "\n",
       "       W_KEIT_KIND_HH  WOHNDAUER_2008       WOHNLAGE       ZABEOTYP  \\\n",
       "count   137910.000000   145056.000000  141725.000000  191652.000000   \n",
       "mean         4.152716        8.646371       3.723133       2.576806   \n",
       "std          1.974375        1.154001       2.095540       1.168486   \n",
       "min          0.000000        1.000000       0.000000       1.000000   \n",
       "25%          2.000000        9.000000       2.000000       1.000000   \n",
       "50%          5.000000        9.000000       3.000000       3.000000   \n",
       "75%          6.000000        9.000000       5.000000       3.000000   \n",
       "max          6.000000        9.000000       8.000000       6.000000   \n",
       "\n",
       "       ONLINE_PURCHASE      ANREDE_KZ  ALTERSKATEGORIE_GROB  \n",
       "count    191652.000000  191652.000000         191652.000000  \n",
       "mean          0.090247       1.376432              3.060907  \n",
       "std           0.286536       0.484492              1.086254  \n",
       "min           0.000000       1.000000              1.000000  \n",
       "25%           0.000000       1.000000              3.000000  \n",
       "50%           0.000000       1.000000              3.000000  \n",
       "75%           0.000000       2.000000              4.000000  \n",
       "max           1.000000       2.000000              9.000000  \n",
       "\n",
       "[8 rows x 362 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers. describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>Attribute</th>\n",
       "      <th>Description</th>\n",
       "      <th>Value</th>\n",
       "      <th>Meaning</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NaN</td>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>best-ager typology</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>no classification possible</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>passive elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>cultural elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>experience-driven elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>NaN</td>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>age classification through prename analysis</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>&lt; 30 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>30 - 45 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>46 - 60 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>&gt; 60 years</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0             Attribute  \\\n",
       "0         NaN              AGER_TYP   \n",
       "1         NaN                   NaN   \n",
       "2         NaN                   NaN   \n",
       "3         NaN                   NaN   \n",
       "4         NaN                   NaN   \n",
       "5         NaN  ALTERSKATEGORIE_GROB   \n",
       "6         NaN                   NaN   \n",
       "7         NaN                   NaN   \n",
       "8         NaN                   NaN   \n",
       "9         NaN                   NaN   \n",
       "\n",
       "                                    Description  Value  \\\n",
       "0                            best-ager typology     -1   \n",
       "1                                           NaN      0   \n",
       "2                                           NaN      1   \n",
       "3                                           NaN      2   \n",
       "4                                           NaN      3   \n",
       "5  age classification through prename analysis   -1, 0   \n",
       "6                                           NaN      1   \n",
       "7                                           NaN      2   \n",
       "8                                           NaN      3   \n",
       "9                                           NaN      4   \n",
       "\n",
       "                      Meaning  \n",
       "0                     unknown  \n",
       "1  no classification possible  \n",
       "2             passive elderly  \n",
       "3            cultural elderly  \n",
       "4   experience-driven elderly  \n",
       "5                     unknown  \n",
       "6                  < 30 years  \n",
       "7               30 - 45 years  \n",
       "8               46 - 60 years  \n",
       "9                  > 60 years  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "attributes_info.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def drop_unnamed_cols(df):\n",
    "    \"\"\"\n",
    "    drop the Unnamed columns columns in the data\n",
    "\n",
    "    \"\"\"\n",
    "    df.drop(columns=['Unnamed: 0'], inplace=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Information level</th>\n",
       "      <th>Attribute</th>\n",
       "      <th>Description</th>\n",
       "      <th>Additional notes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NaN</td>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>best-ager typology</td>\n",
       "      <td>in cooperation with Kantar TNS; the informatio...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Person</td>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>age through prename analysis</td>\n",
       "      <td>modelled on millions of first name-age-referen...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NaN</td>\n",
       "      <td>ANREDE_KZ</td>\n",
       "      <td>gender</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>CJT_GESAMTTYP</td>\n",
       "      <td>Customer-Journey-Typology relating to the pref...</td>\n",
       "      <td>relating to the preferred information, marketi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NaN</td>\n",
       "      <td>FINANZ_MINIMALIST</td>\n",
       "      <td>financial typology: low financial interest</td>\n",
       "      <td>Gfk-Typology based on a representative househo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>308</th>\n",
       "      <td>Community</td>\n",
       "      <td>ARBEIT</td>\n",
       "      <td>share of unemployed person in the community</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>309</th>\n",
       "      <td>NaN</td>\n",
       "      <td>EINWOHNER</td>\n",
       "      <td>inhabitants</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>310</th>\n",
       "      <td>NaN</td>\n",
       "      <td>GKZ</td>\n",
       "      <td>standardized community-code</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>311</th>\n",
       "      <td>NaN</td>\n",
       "      <td>ORTSGR_KLS9</td>\n",
       "      <td>classified number of inhabitants</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>312</th>\n",
       "      <td>NaN</td>\n",
       "      <td>RELAT_AB</td>\n",
       "      <td>share of unemployed in relation to the county ...</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>313 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    Information level             Attribute  \\\n",
       "0                 NaN              AGER_TYP   \n",
       "1              Person  ALTERSKATEGORIE_GROB   \n",
       "2                 NaN             ANREDE_KZ   \n",
       "3                 NaN         CJT_GESAMTTYP   \n",
       "4                 NaN     FINANZ_MINIMALIST   \n",
       "..                ...                   ...   \n",
       "308         Community                ARBEIT   \n",
       "309               NaN             EINWOHNER   \n",
       "310               NaN                   GKZ   \n",
       "311               NaN           ORTSGR_KLS9   \n",
       "312               NaN              RELAT_AB   \n",
       "\n",
       "                                           Description  \\\n",
       "0                                   best-ager typology   \n",
       "1                        age through prename analysis    \n",
       "2                                               gender   \n",
       "3    Customer-Journey-Typology relating to the pref...   \n",
       "4           financial typology: low financial interest   \n",
       "..                                                 ...   \n",
       "308        share of unemployed person in the community   \n",
       "309                                        inhabitants   \n",
       "310                        standardized community-code   \n",
       "311                   classified number of inhabitants   \n",
       "312  share of unemployed in relation to the county ...   \n",
       "\n",
       "                                      Additional notes  \n",
       "0    in cooperation with Kantar TNS; the informatio...  \n",
       "1    modelled on millions of first name-age-referen...  \n",
       "2                                                  NaN  \n",
       "3    relating to the preferred information, marketi...  \n",
       "4    Gfk-Typology based on a representative househo...  \n",
       "..                                                 ...  \n",
       "308                                                NaN  \n",
       "309                                                NaN  \n",
       "310                                                NaN  \n",
       "311                                                NaN  \n",
       "312                                                NaN  \n",
       "\n",
       "[313 rows x 4 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# drop the unnamed columns\n",
    "drop_unnamed_cols(azdias)\n",
    "drop_unnamed_cols(customers)\n",
    "drop_unnamed_cols(attributes_info)\n",
    "drop_unnamed_cols(attributes_des)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Attribute</th>\n",
       "      <th>Description</th>\n",
       "      <th>Value</th>\n",
       "      <th>Meaning</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>best-ager typology</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>no classification possible</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>passive elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>cultural elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>experience-driven elderly</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>age classification through prename analysis</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>&lt; 30 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>30 - 45 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>46 - 60 years</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>&gt; 60 years</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Attribute                                   Description  Value  \\\n",
       "0              AGER_TYP                            best-ager typology     -1   \n",
       "1              AGER_TYP                                           NaN      0   \n",
       "2              AGER_TYP                                           NaN      1   \n",
       "3              AGER_TYP                                           NaN      2   \n",
       "4              AGER_TYP                                           NaN      3   \n",
       "5  ALTERSKATEGORIE_GROB  age classification through prename analysis   -1, 0   \n",
       "6  ALTERSKATEGORIE_GROB                                           NaN      1   \n",
       "7  ALTERSKATEGORIE_GROB                                           NaN      2   \n",
       "8  ALTERSKATEGORIE_GROB                                           NaN      3   \n",
       "9  ALTERSKATEGORIE_GROB                                           NaN      4   \n",
       "\n",
       "                      Meaning  \n",
       "0                     unknown  \n",
       "1  no classification possible  \n",
       "2             passive elderly  \n",
       "3            cultural elderly  \n",
       "4   experience-driven elderly  \n",
       "5                     unknown  \n",
       "6                  < 30 years  \n",
       "7               30 - 45 years  \n",
       "8               46 - 60 years  \n",
       "9                  > 60 years  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fill in the null values in the attribute column\n",
    "attributes_info[\"Attribute\"] = attributes_info[\"Attribute\"].ffill()\n",
    "attributes_info.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Preprocesing "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Address mixed type column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['CAMEO_DEUG_2015', 'CAMEO_INTL_2015'], dtype='object')"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias.columns[18:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['CAMEO_DEUG_2015', 'CAMEO_INTL_2015'], dtype='object')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers.columns[18:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([nan, 51.0, 24.0, 12.0, 43.0, 54.0, 22.0, 14.0, 13.0, 15.0, 33.0,\n",
       "       41.0, 34.0, 55.0, 25.0, 23.0, 31.0, 52.0, 35.0, 45.0, 44.0, 32.0,\n",
       "       '22', '24', '41', '12', '54', '51', '44', '35', '23', '25', '14',\n",
       "       '34', '52', '55', '31', '32', '15', '13', '43', '33', '45', 'XX'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias['CAMEO_INTL_2015'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Attribute</th>\n",
       "      <th>Description</th>\n",
       "      <th>Value</th>\n",
       "      <th>Meaning</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>CAMEO classification 2015 - Uppergroup</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>upper class</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>upper middleclass</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "      <td>established middleclasse</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>consumption-oriented middleclass</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5</td>\n",
       "      <td>active middleclass</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6</td>\n",
       "      <td>low-consumption middleclass</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7</td>\n",
       "      <td>lower middleclass</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8</td>\n",
       "      <td>working class</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>CAMEO_DEUG_2015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9</td>\n",
       "      <td>urban working class</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Attribute                             Description Value  \\\n",
       "51  CAMEO_DEUG_2015  CAMEO classification 2015 - Uppergroup    -1   \n",
       "52  CAMEO_DEUG_2015                                     NaN     1   \n",
       "53  CAMEO_DEUG_2015                                     NaN     2   \n",
       "54  CAMEO_DEUG_2015                                     NaN     3   \n",
       "55  CAMEO_DEUG_2015                                     NaN     4   \n",
       "56  CAMEO_DEUG_2015                                     NaN     5   \n",
       "57  CAMEO_DEUG_2015                                     NaN     6   \n",
       "58  CAMEO_DEUG_2015                                     NaN     7   \n",
       "59  CAMEO_DEUG_2015                                     NaN     8   \n",
       "60  CAMEO_DEUG_2015                                     NaN     9   \n",
       "\n",
       "                              Meaning  \n",
       "51                            unknown  \n",
       "52                        upper class  \n",
       "53                  upper middleclass  \n",
       "54           established middleclasse  \n",
       "55  consumption-oriented middleclass   \n",
       "56                active middleclass   \n",
       "57        low-consumption middleclass  \n",
       "58                  lower middleclass  \n",
       "59                      working class  \n",
       "60                urban working class  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "attributes_info[attributes_info[\"Attribute\"] == \"CAMEO_DEUG_2015\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace the X and XX with nan\n",
    "azdias[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']]=azdias[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']].replace(['X','XX'],-1)\n",
    "azdias[['CAMEO_DEUG_2015','CAMEO_INTL_2015']]=azdias[['CAMEO_DEUG_2015','CAMEO_INTL_2015']].apply(pd.to_numeric)\n",
    "\n",
    "customers[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']]=customers[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']].replace(['X','XX'],-1)\n",
    "customers[['CAMEO_DEUG_2015','CAMEO_INTL_2015']]=customers[['CAMEO_DEUG_2015','CAMEO_INTL_2015']].apply(pd.to_numeric)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Look at the extra columnns in customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers['PRODUCT_GROUP'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers['ONLINE_PURCHASE'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers['CUSTOMER_GROUP'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop the additional columns\n",
    "customers.drop(['PRODUCT_GROUP', 'ONLINE_PURCHASE', 'CUSTOMER_GROUP'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(891221, 310)\n",
      "(191652, 310)\n"
     ]
    }
   ],
   "source": [
    "print(azdias.shape)\n",
    "print(customers.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Looking at the missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                attribute                                 missing_or_unknown\n",
      "0                AGER_TYP                                   [-1, 0, 1, 2, 3]\n",
      "1    ALTERSKATEGORIE_GROB                            [-1,  0, 1, 2, 3, 4, 9]\n",
      "2                ALTER_HH  [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,...\n",
      "3               ANREDE_KZ                                     [-1,  0, 1, 2]\n",
      "4     ANZ_HAUSHALTE_AKTIV                                                […]\n",
      "..                    ...                                                ...\n",
      "309        WOHNDAUER_2008                [-1,  0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\n",
      "310              WOHNLAGE                       [-1, 0, 1, 2, 3, 4, 5, 7, 8]\n",
      "311    WACHSTUMSGEBIET_NB                            [-1,  0, 1, 2, 3, 4, 5]\n",
      "312        W_KEIT_KIND_HH                         [-1,  0, 1, 2, 3, 4, 5, 6]\n",
      "313              ZABEOTYP                         [-1,  9, 1, 2, 3, 4, 5, 6]\n",
      "\n",
      "[314 rows x 2 columns]\n"
     ]
    }
   ],
   "source": [
    "# Attributes with missing or unknown\n",
    "unknowns = []\n",
    "for attribute in attributes_info['Attribute'].unique():\n",
    "    _ = attributes_info.loc[attributes_info['Attribute'] == attribute, 'Value'].astype(str).str.cat(sep=',')\n",
    "    _ = _.split(',')\n",
    "    unknowns.append(_)\n",
    "    \n",
    "unknowns = pd.concat([pd.Series(attributes_info['Attribute'].unique()), pd.Series(unknowns)], axis=1)\n",
    "unknowns.columns = ['attribute', 'missing_or_unknown']\n",
    "print(unknowns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>column</th>\n",
       "      <th>azdias_nan</th>\n",
       "      <th>azdiaz_nan_percent</th>\n",
       "      <th>customers_nan</th>\n",
       "      <th>customer_nan_percent</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ALTER_KIND4</td>\n",
       "      <td>890016</td>\n",
       "      <td>99.864792</td>\n",
       "      <td>191416</td>\n",
       "      <td>99.876860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>ALTER_KIND3</td>\n",
       "      <td>885051</td>\n",
       "      <td>99.307691</td>\n",
       "      <td>190377</td>\n",
       "      <td>99.334732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ALTER_KIND2</td>\n",
       "      <td>861722</td>\n",
       "      <td>96.690047</td>\n",
       "      <td>186552</td>\n",
       "      <td>97.338927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ALTER_KIND1</td>\n",
       "      <td>810163</td>\n",
       "      <td>90.904837</td>\n",
       "      <td>179886</td>\n",
       "      <td>93.860748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>EXTSEL992</td>\n",
       "      <td>654153</td>\n",
       "      <td>73.399639</td>\n",
       "      <td>85283</td>\n",
       "      <td>44.498883</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>346</th>\n",
       "      <td>SHOPPER_TYP</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354</th>\n",
       "      <td>VERS_TYP</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>363</th>\n",
       "      <td>ZABEOTYP</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>364</th>\n",
       "      <td>ANREDE_KZ</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>365</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>366 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                   column  azdias_nan  azdiaz_nan_percent  customers_nan  \\\n",
       "7             ALTER_KIND4      890016           99.864792         191416   \n",
       "6             ALTER_KIND3      885051           99.307691         190377   \n",
       "5             ALTER_KIND2      861722           96.690047         186552   \n",
       "4             ALTER_KIND1      810163           90.904837         179886   \n",
       "100             EXTSEL992      654153           73.399639          85283   \n",
       "..                    ...         ...                 ...            ...   \n",
       "346           SHOPPER_TYP           0            0.000000              0   \n",
       "354              VERS_TYP           0            0.000000              0   \n",
       "363              ZABEOTYP           0            0.000000              0   \n",
       "364             ANREDE_KZ           0            0.000000              0   \n",
       "365  ALTERSKATEGORIE_GROB           0            0.000000              0   \n",
       "\n",
       "     customer_nan_percent  \n",
       "7               99.876860  \n",
       "6               99.334732  \n",
       "5               97.338927  \n",
       "4               93.860748  \n",
       "100             44.498883  \n",
       "..                    ...  \n",
       "346              0.000000  \n",
       "354              0.000000  \n",
       "363              0.000000  \n",
       "364              0.000000  \n",
       "365              0.000000  \n",
       "\n",
       "[366 rows x 5 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# creating a dataframe to get count/percentage of missing values in azdias and customers\n",
    "azdias_nan=pd.DataFrame(azdias.isna().sum(axis=0)).reset_index()\n",
    "azdias_nan.columns=['column','azdias_nan']\n",
    "azdias_nan['azdiaz_nan_percent']=(azdias_nan['azdias_nan']/len(azdias))*100\n",
    "\n",
    "customers_nan=pd.DataFrame(customers.isna().sum(axis=0)).reset_index()\n",
    "customers_nan.columns=['column','customers_nan']\n",
    "customers_nan['customer_nan_percent']=(customers_nan['customers_nan']/len(customers))*100\n",
    "missing_values=pd.merge(azdias_nan, customers_nan, on='column')\n",
    "missing_values=missing_values.sort_values(by=['azdias_nan','customers_nan'],ascending=[False,False])\n",
    "missing_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,8))\n",
    "# set width of bar\n",
    "barWidth = 0.25\n",
    "# set height of bar\n",
    "bars1 = missing_values['azdiaz_nan_percent'][:10]\n",
    "bars2 = missing_values['customer_nan_percent'][:10]\n",
    "label=missing_values['column'][:10]\n",
    "# Set position of bar on X axis\n",
    "r1 = np.arange(len(bars1))\n",
    "r2 = [x + barWidth for x in r1]\n",
    " \n",
    "# Make the plot\n",
    "plt.bar(r1, bars1, width=barWidth, edgecolor='white')\n",
    "plt.bar(r2, bars2, width=barWidth, edgecolor='white')\n",
    "\n",
    " \n",
    "# Add xticks on the middle of the group bars\n",
    "plt.xlabel('Features')\n",
    "plt.ylabel('Percentage')\n",
    "plt.xticks([r + barWidth/2 for r in range(len(bars1))], label,rotation=90)\n",
    " \n",
    "# Create legend & Show graphic\n",
    "plt.title('Top 10 Features with most missing values')\n",
    "plt.legend(['azdiaz_nan_percent','customer_nan_percent'],prop={'size': 12})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ALTER_KIND4', 'ALTER_KIND3', 'ALTER_KIND2', 'ALTER_KIND1']"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we will drop the columns with highest missing value\n",
    "# Only the first variable has more than 60% missing value\n",
    "drop_cols=list(label[:4])\n",
    "drop_cols"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of Numeric columns 360\n",
      "Length of categorical columns 6\n",
      "Categorial columns are {'D19_LETZTER_KAUF_BRANCHE', 'EINGEFUEGT_AM', 'OST_WEST_KZ', 'CAMEO_INTL_2015', 'CAMEO_DEU_2015', 'CAMEO_DEUG_2015'}\n"
     ]
    }
   ],
   "source": [
    "# Getting all numeric columns and categorical columns\n",
    "Numeric_columns=azdias.select_dtypes(include=np.number).columns.tolist()\n",
    "categorical_col=set(azdias.columns).difference(set(Numeric_columns))\n",
    "print(\"Length of Numeric columns\",len(Numeric_columns))\n",
    "print(\"Length of categorical columns\",len(categorical_col))\n",
    "print(\"Categorial columns are\",categorical_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert variables to numeric types\n",
    "azdias[Numeric_columns] = azdias[Numeric_columns].apply(pd.to_numeric)\n",
    "customers[Numeric_columns] = customers[Numeric_columns].apply(pd.to_numeric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'D19_LETZTER_KAUF_BRANCHE': array(['D19_UNBEKANNT', 'D19_SCHUHE', 'D19_ENERGIE', 'D19_KOSMETIK',\n",
       "        'D19_VOLLSORTIMENT', 'D19_SONSTIGE', 'D19_BANKEN_GROSS',\n",
       "        'D19_DROGERIEARTIKEL', 'D19_HANDWERK', 'D19_BUCH_CD',\n",
       "        'D19_VERSICHERUNGEN', 'D19_VERSAND_REST', 'D19_TELKO_REST',\n",
       "        'D19_BANKEN_DIREKT', 'D19_BANKEN_REST', 'D19_FREIZEIT',\n",
       "        'D19_LEBENSMITTEL', 'D19_HAUS_DEKO', 'D19_BEKLEIDUNG_REST',\n",
       "        'D19_SAMMELARTIKEL', 'D19_TELKO_MOBILE', 'D19_REISEN',\n",
       "        'D19_BEKLEIDUNG_GEH', 'D19_TECHNIK', 'D19_NAHRUNGSERGAENZUNG',\n",
       "        'D19_DIGIT_SERV', 'D19_LOTTO', 'D19_RATGEBER', 'D19_TIERARTIKEL',\n",
       "        'D19_KINDERARTIKEL', 'D19_BIO_OEKO', 'D19_WEIN_FEINKOST',\n",
       "        'D19_GARTEN', 'D19_BILDUNG', 'D19_BANKEN_LOKAL'], dtype=object),\n",
       " 'EINGEFUEGT_AM': array(['1992-02-10 00:00:00', '1992-02-12 00:00:00',\n",
       "        '1997-04-21 00:00:00', ..., '2010-12-02 00:00:00',\n",
       "        '2005-03-19 00:00:00', '2011-11-18 00:00:00'], dtype=object),\n",
       " 'OST_WEST_KZ': array(['W', 'O'], dtype=object),\n",
       " 'CAMEO_INTL_2015': array([51.0, 24.0, 12.0, 43.0, 54.0, 22.0, 14.0, 13.0, 15.0, 33.0, 41.0,\n",
       "        34.0, 55.0, 25.0, 23.0, 31.0, 52.0, 35.0, 45.0, 44.0, 32.0, '22',\n",
       "        '24', '41', '12', '54', '51', '44', '35', '23', '25', '14', '34',\n",
       "        '52', '55', '31', '32', '15', '13', '43', '33', '45', 'XX'],\n",
       "       dtype=object),\n",
       " 'CAMEO_DEU_2015': array(['8A', '4C', '2A', '6B', '8C', '4A', '2D', '1A', '1E', '9D', '5C',\n",
       "        '8B', '7A', '5D', '9E', '9B', '1B', '3D', '4E', '4B', '3C', '5A',\n",
       "        '7B', '9A', '6D', '6E', '2C', '7C', '9C', '7D', '5E', '1D', '8D',\n",
       "        '6C', '6A', '5B', '4D', '3A', '2B', '7E', '3B', '6F', '5F', '1C',\n",
       "        'XX'], dtype=object),\n",
       " 'CAMEO_DEUG_2015': array([8.0, 4.0, 2.0, 6.0, 1.0, 9.0, 5.0, 7.0, 3.0, '4', '3', '7', '2',\n",
       "        '8', '9', '6', '5', '1', 'X'], dtype=object)}"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Further analysis of categorical columns\n",
    "category={}\n",
    "def categorical_check():\n",
    "    '''\n",
    "    return a list of unique categorical variables \n",
    "    \n",
    "    '''\n",
    "    for i in categorical_col:\n",
    "        category[i]=azdias[i].dropna().unique()\n",
    "    return category\n",
    "categorical_check()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 'EINGEFUEGT_AM' is time data\n",
    "- 'CAMEO_INTL_2015', 'CAMEO_DEU_2015' and 'CAMEO_DEUG_2015' have XX and X as values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# After further analysis, these variables will be dropped\n",
    "azdias=azdias.drop('D19_LETZTER_KAUF_BRANCHE',axis=1)\n",
    "customers=customers.drop('D19_LETZTER_KAUF_BRANCHE',axis=1)\n",
    "\n",
    "azdias=azdias.drop('EINGEFUEGT_AM',axis=1)\n",
    "customers=customers.drop('EINGEFUEGT_AM',axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Further analysis of missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "unknown                            232\n",
       "average                            162\n",
       "low                                154\n",
       "high                               154\n",
       "very high                          142\n",
       "very low                           139\n",
       "none                               138\n",
       "Multibuyer 0-12 months              34\n",
       "no transaction known                34\n",
       "Prospects > 24 months               34\n",
       "Singlebuyer 13-24 months            34\n",
       "Doublebuyer 0-12 months             34\n",
       "Singlebuyer 0-12 months             34\n",
       "Buyer > 24 months                   34\n",
       "Multi-/Doublebuyer 13-24 months     34\n",
       "no transactions known               22\n",
       "highest affinity                    14\n",
       "low affinity                        14\n",
       "lowest affinity                     14\n",
       "very high affinity                  14\n",
       "Name: Meaning, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's have a closer look at 'Meaning' column in attribute values\n",
    "attributes_info['Meaning'].value_counts()[:20]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "we have lots of unknown value in our dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Attribute</th>\n",
       "      <th>Description</th>\n",
       "      <th>Value</th>\n",
       "      <th>Meaning</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AGER_TYP</td>\n",
       "      <td>best-ager typology</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ALTERSKATEGORIE_GROB</td>\n",
       "      <td>age classification through prename analysis</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>ANREDE_KZ</td>\n",
       "      <td>gender</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>BALLRAUM</td>\n",
       "      <td>distance to next urban centre</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>BIP_FLAG</td>\n",
       "      <td>business-flag indicating companies in the buil...</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2219</th>\n",
       "      <td>WOHNDAUER_2008</td>\n",
       "      <td>length of residence</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2229</th>\n",
       "      <td>WOHNLAGE</td>\n",
       "      <td>residential-area</td>\n",
       "      <td>-1</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2238</th>\n",
       "      <td>WACHSTUMSGEBIET_NB</td>\n",
       "      <td>growing area (population growth in the last 5 ...</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2244</th>\n",
       "      <td>W_KEIT_KIND_HH</td>\n",
       "      <td>likelihood of a child present in this household</td>\n",
       "      <td>-1, 0</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2251</th>\n",
       "      <td>ZABEOTYP</td>\n",
       "      <td>typification of energy consumers</td>\n",
       "      <td>-1, 9</td>\n",
       "      <td>unknown</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>232 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Attribute                                        Description  \\\n",
       "0                 AGER_TYP                                 best-ager typology   \n",
       "5     ALTERSKATEGORIE_GROB       age classification through prename analysis    \n",
       "33               ANREDE_KZ                                             gender   \n",
       "40                BALLRAUM                     distance to next urban centre    \n",
       "48                BIP_FLAG  business-flag indicating companies in the buil...   \n",
       "...                    ...                                                ...   \n",
       "2219        WOHNDAUER_2008                                length of residence   \n",
       "2229              WOHNLAGE                                  residential-area    \n",
       "2238    WACHSTUMSGEBIET_NB  growing area (population growth in the last 5 ...   \n",
       "2244        W_KEIT_KIND_HH    likelihood of a child present in this household   \n",
       "2251              ZABEOTYP                   typification of energy consumers   \n",
       "\n",
       "      Value  Meaning  \n",
       "0        -1  unknown  \n",
       "5     -1, 0  unknown  \n",
       "33    -1, 0  unknown  \n",
       "40       -1  unknown  \n",
       "48       -1  unknown  \n",
       "...     ...      ...  \n",
       "2219  -1, 0  unknown  \n",
       "2229     -1  unknown  \n",
       "2238  -1, 0  unknown  \n",
       "2244  -1, 0  unknown  \n",
       "2251  -1, 9  unknown  \n",
       "\n",
       "[232 rows x 4 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Investigate the unknown values\n",
    "unknown_attributes_info = attributes_info[attributes_info[\"Meaning\"] == \"unknown\"]\n",
    "unknown_attributes_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The unknown values are not consistent throughtout the dataset\n",
    "- Some are represented as (-1,0) or (-1,9) or just -1. \n",
    "- We need to fix this and make the data consistent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make data consistent\n",
    "fix_unknown1=list(set(list(attributes_info[attributes_info['Value']=='-1, 0']['Attribute'])).intersection(set(azdias.columns)))\n",
    "fix_unknown9=list(set(list(attributes_info[attributes_info['Value']=='-1, 9']['Attribute'])).intersection(set(azdias.columns)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['KKK', 'KBA05_GBZ', 'PRAEGENDE_JUGENDJAHRE', 'TITEL_KZ', 'HH_EINKOMMEN_SCORE', 'ANREDE_KZ', 'REGIOTYP', 'WOHNDAUER_2008', 'W_KEIT_KIND_HH', 'KBA05_BAUMAX', 'GEBAEUDETYP', 'NATIONALITAET_KZ', 'ALTERSKATEGORIE_GROB']\n"
     ]
    }
   ],
   "source": [
    "print(fix_unknown1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['KBA05_ALTER1', 'KBA05_HERST2', 'SEMIO_FAM', 'KBA05_KRSAQUOT', 'KBA05_MOD1', 'KBA05_HERST1', 'SEMIO_MAT', 'SEMIO_TRADV', 'KBA05_AUTOQUOT', 'KBA05_MAXHERST', 'KBA05_DIESEL', 'KBA05_KRSOBER', 'KBA05_ZUL3', 'SEMIO_KAEM', 'SEMIO_VERT', 'KBA05_CCM3', 'KBA05_ALTER4', 'SEMIO_REL', 'KBA05_KW3', 'KBA05_SEG3', 'KBA05_HERST4', 'KBA05_ZUL4', 'KBA05_MAXAH', 'KBA05_KRSHERST3', 'KBA05_KRSHERST1', 'SEMIO_PFLICHT', 'KBA05_KRSVAN', 'KBA05_MOTRAD', 'KBA05_HERST5', 'KBA05_ANHANG', 'KBA05_ZUL1', 'KBA05_CCM2', 'SEMIO_RAT', 'KBA05_VORB0', 'KBA05_ALTER2', 'KBA05_SEG4', 'KBA05_MAXBJ', 'KBA05_MOD8', 'KBA05_KRSKLEIN', 'KBA05_SEG9', 'SEMIO_ERL', 'SEMIO_KRIT', 'KBA05_SEG5', 'KBA05_KW2', 'KBA05_SEG7', 'KBA05_MAXSEG', 'KBA05_HERSTTEMP', 'KBA05_ALTER3', 'KBA05_KRSHERST2', 'KBA05_KW1', 'KBA05_SEG10', 'KBA05_MOD2', 'KBA05_SEG2', 'KBA05_SEG8', 'KBA05_VORB1', 'KBA05_MOTOR', 'KBA05_KRSZUL', 'SEMIO_SOZ', 'RELAT_AB', 'KBA05_MAXVORB', 'KBA05_FRAU', 'KBA05_MOD4', 'SEMIO_DOM', 'KBA05_HERST3', 'KBA05_MOD3', 'KBA05_SEG6', 'SEMIO_LUST', 'KBA05_ZUL2', 'KBA05_CCM1', 'KBA05_MODTEMP', 'ZABEOTYP', 'KBA05_SEG1', 'SEMIO_KULT', 'KBA05_CCM4', 'KBA05_VORB2']\n"
     ]
    }
   ],
   "source": [
    "print(fix_unknown9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "azdias[fix_unknown1]=azdias[fix_unknown1].replace(0,-1)\n",
    "azdias[fix_unknown9]=azdias[fix_unknown9].replace(9,-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers[fix_unknown1]=customers[fix_unknown1].replace(0,-1)\n",
    "customers[fix_unknown9]=customers[fix_unknown9].replace(9,-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Missing value in rows and columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remove rows with most missing values\n",
    "azdias=azdias[azdias.isnull().sum(axis=1)<=20]\n",
    "\n",
    "customers=customers[customers.isnull().sum(axis=1)<=20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dropping columns with most missing values\n",
    "azdias=azdias.drop(drop_cols,axis=1)\n",
    "\n",
    "customers=customers.drop(drop_cols,axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Drop Correlated features above threshold of 75%\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# find correlation matrix and drop variables with greater than 75% correlation  \n",
    "corr_matrix = azdias.corr().abs()\n",
    "upper_limit = corr_matrix.where(np.triu(np.ones(corr_matrix.shape), k=1).astype(np.bool))\n",
    "# identify columns to drop based on threshold limit\n",
    "drop_columns = [column for column in upper_limit.columns if any(upper_limit[column] > .75)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "azdias.drop(drop_columns,axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers.drop(drop_columns,axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imputing missing values\n",
    "Find and relace nan values witht the mode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2664489\n",
      "131091\n"
     ]
    }
   ],
   "source": [
    "print(azdias.isna().sum().sum())\n",
    "print(customers.isna().sum().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Percent of missing values')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#looking for nans in columns\n",
    "column_nans = azdias.isnull().mean()\n",
    "plt.hist(column_nans, bins = np.arange(0,1+.05,.05))\n",
    "plt.ylabel('features')\n",
    "plt.xlabel('Percent of missing values')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "azdias=azdias.fillna(azdias.mode().iloc[0])\n",
    "customers=customers.fillna(customers.mode().iloc[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "print(azdias.isna().sum().sum())\n",
    "print(customers.isna().sum().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(745305, 360)\n",
      "(132095, 363)\n"
     ]
    }
   ],
   "source": [
    "print(azdias.shape)\n",
    "print(customers.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### One-hot encoding categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of Numeric columns in azdias 358\n",
      "Length of categorical columns in azdias 2\n",
      "Categorial columns in azdias are {'OST_WEST_KZ', 'CAMEO_DEU_2015'}\n"
     ]
    }
   ],
   "source": [
    "# Getting all numeric columns and categorical columns\n",
    "Numeric_columns=azdias.select_dtypes(include=np.number).columns.tolist()\n",
    "categorical_col=set(azdias.columns).difference(set(Numeric_columns))\n",
    "print(\"Length of Numeric columns in azdias\",len(Numeric_columns))\n",
    "print(\"Length of categorical columns in azdias\",len(categorical_col))\n",
    "print(\"Categorial columns in azdias are\",categorical_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of Numeric columns in customers 359\n",
      "Length of categorical columns in customers 4\n",
      "Categorial columns in customers are {'PRODUCT_GROUP', 'OST_WEST_KZ', 'CAMEO_DEU_2015', 'CUSTOMER_GROUP'}\n"
     ]
    }
   ],
   "source": [
    "# Getting all numeric columns and categorical columns\n",
    "Numeric_columns=customers.select_dtypes(include=np.number).columns.tolist()\n",
    "categorical_col=set(customers.columns).difference(set(Numeric_columns))\n",
    "print(\"Length of Numeric columns in customers\",len(Numeric_columns))\n",
    "print(\"Length of categorical columns in customers\",len(categorical_col))\n",
    "print(\"Categorial columns in customers are\",categorical_col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "azdias=pd.get_dummies(azdias)\n",
    "customers=pd.get_dummies(customers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(745305, 405)\n",
      "(132095, 411)\n"
     ]
    }
   ],
   "source": [
    "print(azdias.shape)\n",
    "print(customers.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Standardize the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "azdias[azdias.columns] = scaler.fit_transform(azdias)\n",
    "customers[customers.columns] = scaler.fit_transform(customers)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Complete data preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "azdias = pd.read_csv('Udacity_AZDIAS_reduce_mem_052018.csv')\n",
    "customers = pd.read_csv('Udacity_CUSTOMERS_reduce_mem_052018.csv')\n",
    "def drop_unnamed_cols(df):\n",
    "    \"\"\"\n",
    "    drop the Unnamed columns in the data\n",
    "\n",
    "    \"\"\"\n",
    "    df.drop(columns=['Unnamed: 0'], inplace=True)\n",
    "    return df\n",
    "\n",
    "fix_unknown1 = ['KKK', 'KBA05_GBZ', 'PRAEGENDE_JUGENDJAHRE', 'TITEL_KZ', 'HH_EINKOMMEN_SCORE', 'ANREDE_KZ', 'REGIOTYP', \n",
    "                'WOHNDAUER_2008', 'W_KEIT_KIND_HH', 'KBA05_BAUMAX', 'GEBAEUDETYP', 'NATIONALITAET_KZ', 'ALTERSKATEGORIE_GROB']\n",
    "\n",
    "fix_unknown9 =['KBA05_ALTER1', 'KBA05_HERST2', 'SEMIO_FAM', 'KBA05_KRSAQUOT', 'KBA05_MOD1', 'KBA05_HERST1', 'SEMIO_MAT', \n",
    "                'SEMIO_TRADV', 'KBA05_AUTOQUOT', 'KBA05_MAXHERST', 'KBA05_DIESEL', 'KBA05_KRSOBER', 'KBA05_ZUL3', 'SEMIO_KAEM',\n",
    "               'SEMIO_VERT', 'KBA05_CCM3', 'KBA05_ALTER4', 'SEMIO_REL', 'KBA05_KW3', 'KBA05_SEG3', 'KBA05_HERST4', \n",
    "               'KBA05_ZUL4', 'KBA05_MAXAH', 'KBA05_KRSHERST3', 'KBA05_KRSHERST1', 'SEMIO_PFLICHT', 'KBA05_KRSVAN', \n",
    "               'KBA05_MOTRAD', 'KBA05_HERST5', 'KBA05_ANHANG', 'KBA05_ZUL1', 'KBA05_CCM2', 'SEMIO_RAT', 'KBA05_VORB0', \n",
    "               'KBA05_ALTER2', 'KBA05_SEG4', 'KBA05_MAXBJ', 'KBA05_MOD8', 'KBA05_KRSKLEIN', 'KBA05_SEG9', 'SEMIO_ERL', \n",
    "               'SEMIO_KRIT', 'KBA05_SEG5', 'KBA05_KW2', 'KBA05_SEG7', 'KBA05_MAXSEG', 'KBA05_HERSTTEMP', 'KBA05_ALTER3', \n",
    "               'KBA05_KRSHERST2', 'KBA05_KW1', 'KBA05_SEG10', 'KBA05_MOD2', 'KBA05_SEG2', 'KBA05_SEG8', 'KBA05_VORB1', \n",
    "               'KBA05_MOTOR', 'KBA05_KRSZUL', 'SEMIO_SOZ', 'RELAT_AB', 'KBA05_MAXVORB', 'KBA05_FRAU', 'KBA05_MOD4', \n",
    "               'SEMIO_DOM', 'KBA05_HERST3', 'KBA05_MOD3', 'KBA05_SEG6', 'SEMIO_LUST', 'KBA05_ZUL2', 'KBA05_CCM1', \n",
    "               'KBA05_MODTEMP', 'ZABEOTYP', 'KBA05_SEG1', 'SEMIO_KULT', 'KBA05_CCM4', 'KBA05_VORB2']\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "def data_cleaning(df,name=None):\n",
    "    '''\n",
    "    Function take data as uinput and pre_processed the data \n",
    "    Output is cleaned data\n",
    "    \n",
    "    '''\n",
    "#     drop_unnamed_cols(df)\n",
    "    print(\"Shape before\",df.shape)\n",
    "    if name=='customers':\n",
    "        df=df.drop(['PRODUCT_GROUP','CUSTOMER_GROUP','ONLINE_PURCHASE'],axis=1)\n",
    "    \n",
    "    if name=='azdias':\n",
    "        df=df[df.isnull().sum(axis=1)<=20].reset_index(drop=True)\n",
    "        \n",
    "    # finding numeric and categorical columns\n",
    "    Numeric_columns=df.select_dtypes(include=np.number).columns.tolist()\n",
    "    categorical_col=set(df.columns).difference(set(Numeric_columns))\n",
    "    \n",
    "    # numeric cols to numeric\n",
    "    print(categorical_col)\n",
    "    df[Numeric_columns]=df[Numeric_columns].apply(pd.to_numeric)\n",
    "    \n",
    "    # drop columns to avoid lot of columns\n",
    "    df=df.drop('D19_LETZTER_KAUF_BRANCHE',axis=1)\n",
    "    df=df.drop('EINGEFUEGT_AM',axis=1)\n",
    "    df=df.drop(['ALTER_KIND4', 'ALTER_KIND3', 'ALTER_KIND2', 'ALTER_KIND1'],axis=1)\n",
    "\n",
    "    \n",
    "    df[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']]=df[['CAMEO_DEUG_2015','CAMEO_INTL_2015','CAMEO_DEU_2015']].replace(['X','XX'],-1)\n",
    "    df[['CAMEO_DEUG_2015','CAMEO_INTL_2015']]=df[['CAMEO_DEUG_2015','CAMEO_INTL_2015']].apply(pd.to_numeric)\n",
    "    #\n",
    "\n",
    "    df[fix_unknown1]=df[fix_unknown1].replace(0,-1)\n",
    "    df[fix_unknown9]=df[fix_unknown9].replace(9,-1)\n",
    "    \n",
    "    # find correlation matrix    \n",
    "    corr_matrix = df.corr().abs()\n",
    "    upper_limit = corr_matrix.where(np.triu(np.ones(corr_matrix.shape), k=1).astype(np.bool))\n",
    "    # identify columns to drop based on threshold limit\n",
    "    drop_columns = [column for column in upper_limit.columns if any(upper_limit[column] > .75)]\n",
    "    # drop columns from df\n",
    "    df = df.drop(drop_columns, axis=1)\n",
    "    print('shape after corr', df.shape)\n",
    "    \n",
    "    # fill nan with mode\n",
    "    df=df.fillna(df.mode().iloc[0])\n",
    "    print(\"Number of nan Values\",df.isna().sum().sum())\n",
    "    df=pd.get_dummies(df)\n",
    "    \n",
    "    # Standardised data\n",
    "    scaler = StandardScaler()\n",
    "    df[df.columns] = scaler.fit_transform(df)\n",
    "    print(\"shape after\",df.shape)\n",
    "    df = df.set_index('LNR')\n",
    "    \n",
    "    return df\n",
    "\n",
    "# azdias_clean=data_cleaning(azdias,'azdias')\n",
    "# customers_clean=data_cleaning(customers,'customers')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Fix columns difference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(744305, 323)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias_clean.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(191652, 320)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_clean.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'D19_KONSUMTYP', 'ANREDE_KZ', 'KBA13_BAUMAX', 'GEBAEUDETYP_RASTER', 'VK_DHT4A', 'CJT_TYP_3', 'WOHNDAUER_2008', 'KONSUMZELLE'}\n"
     ]
    }
   ],
   "source": [
    "# Difference in columns\n",
    "print(set(list(customers_clean.columns)) - set(list(azdias_clean.columns)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'VERS_TYP', 'FINANZ_UNAUFFAELLIGER', 'NATIONALITAET_KZ', 'SEMIO_REL', 'KBA13_BJ_2000', 'KOMBIALTER', 'SEMIO_KRIT', 'SEMIO_KAEM', 'FINANZ_ANLEGER', 'CJT_TYP_4', 'KBA13_HHZ'}\n"
     ]
    }
   ],
   "source": [
    "# Difference in columns\n",
    "print(set(list(azdias_clean.columns)) - set(list(customers_clean.columns)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop columns\n",
    "customers_clean.drop(columns=['GEBAEUDETYP_RASTER', 'D19_KONSUMTYP', 'KBA13_BAUMAX', 'KONSUMZELLE', 'CJT_TYP_3', 'ANREDE_KZ',\n",
    "                              'WOHNDAUER_2008', 'VK_DHT4A'], inplace=True)\n",
    "\n",
    "azdias_clean.drop(columns=['SEMIO_REL', 'FINANZ_ANLEGER', 'KBA13_HHZ', 'KOMBIALTER', 'KBA13_BJ_2000', 'CJT_TYP_4', 'SEMIO_KAEM',\n",
    "                           'SEMIO_KRIT', 'FINANZ_UNAUFFAELLIGER', 'NATIONALITAET_KZ', 'VERS_TYP'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(744305, 312)\n",
      "(191652, 312)\n"
     ]
    }
   ],
   "source": [
    "print(azdias_clean.shape)\n",
    "print(customers_clean.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>ANZ_HH_TITEL</th>\n",
       "      <th>ANZ_KINDER</th>\n",
       "      <th>ANZ_PERSONEN</th>\n",
       "      <th>ANZ_TITEL</th>\n",
       "      <th>ARBEIT</th>\n",
       "      <th>...</th>\n",
       "      <th>CAMEO_DEU_2015_8B</th>\n",
       "      <th>CAMEO_DEU_2015_8C</th>\n",
       "      <th>CAMEO_DEU_2015_8D</th>\n",
       "      <th>CAMEO_DEU_2015_9A</th>\n",
       "      <th>CAMEO_DEU_2015_9B</th>\n",
       "      <th>CAMEO_DEU_2015_9C</th>\n",
       "      <th>CAMEO_DEU_2015_9D</th>\n",
       "      <th>CAMEO_DEU_2015_9E</th>\n",
       "      <th>OST_WEST_KZ_O</th>\n",
       "      <th>OST_WEST_KZ_W</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "      <td>7.443050e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>8.343793e-15</td>\n",
       "      <td>-1.123348e-14</td>\n",
       "      <td>-2.371110e-15</td>\n",
       "      <td>4.843493e-16</td>\n",
       "      <td>5.360674e-15</td>\n",
       "      <td>1.091576e-14</td>\n",
       "      <td>4.510088e-15</td>\n",
       "      <td>-1.590556e-14</td>\n",
       "      <td>-2.702642e-15</td>\n",
       "      <td>-1.852886e-15</td>\n",
       "      <td>...</td>\n",
       "      <td>-5.352440e-15</td>\n",
       "      <td>1.058515e-14</td>\n",
       "      <td>2.018845e-14</td>\n",
       "      <td>1.703655e-15</td>\n",
       "      <td>-3.483268e-14</td>\n",
       "      <td>-3.938322e-15</td>\n",
       "      <td>2.873153e-15</td>\n",
       "      <td>9.391989e-15</td>\n",
       "      <td>2.676556e-14</td>\n",
       "      <td>-2.676556e-14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "      <td>1.000001e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-5.731380e-01</td>\n",
       "      <td>-9.423887e-01</td>\n",
       "      <td>-1.419539e+00</td>\n",
       "      <td>-3.087966e+00</td>\n",
       "      <td>-5.435379e-01</td>\n",
       "      <td>-1.285939e-01</td>\n",
       "      <td>-2.970230e-01</td>\n",
       "      <td>-1.486979e+00</td>\n",
       "      <td>-6.049438e-02</td>\n",
       "      <td>-2.200629e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>-2.118226e-01</td>\n",
       "      <td>-2.056151e-01</td>\n",
       "      <td>-1.533955e-01</td>\n",
       "      <td>-1.663754e-01</td>\n",
       "      <td>-1.933705e-01</td>\n",
       "      <td>-1.834139e-01</td>\n",
       "      <td>-1.964120e-01</td>\n",
       "      <td>-8.703665e-02</td>\n",
       "      <td>-5.198924e-01</td>\n",
       "      <td>-1.923475e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-5.731380e-01</td>\n",
       "      <td>-9.423887e-01</td>\n",
       "      <td>-1.419539e+00</td>\n",
       "      <td>-4.315714e-01</td>\n",
       "      <td>-4.159390e-01</td>\n",
       "      <td>-1.285939e-01</td>\n",
       "      <td>-2.970230e-01</td>\n",
       "      <td>-6.270170e-01</td>\n",
       "      <td>-6.049438e-02</td>\n",
       "      <td>-1.898515e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-2.118226e-01</td>\n",
       "      <td>-2.056151e-01</td>\n",
       "      <td>-1.533955e-01</td>\n",
       "      <td>-1.663754e-01</td>\n",
       "      <td>-1.933705e-01</td>\n",
       "      <td>-1.834139e-01</td>\n",
       "      <td>-1.964120e-01</td>\n",
       "      <td>-8.703665e-02</td>\n",
       "      <td>-5.198924e-01</td>\n",
       "      <td>5.198924e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-5.731380e-01</td>\n",
       "      <td>-3.943110e-01</td>\n",
       "      <td>2.840275e-01</td>\n",
       "      <td>2.325272e-01</td>\n",
       "      <td>-2.883401e-01</td>\n",
       "      <td>-1.285939e-01</td>\n",
       "      <td>-2.970230e-01</td>\n",
       "      <td>-6.270170e-01</td>\n",
       "      <td>-6.049438e-02</td>\n",
       "      <td>-1.898515e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-2.118226e-01</td>\n",
       "      <td>-2.056151e-01</td>\n",
       "      <td>-1.533955e-01</td>\n",
       "      <td>-1.663754e-01</td>\n",
       "      <td>-1.933705e-01</td>\n",
       "      <td>-1.834139e-01</td>\n",
       "      <td>-1.964120e-01</td>\n",
       "      <td>-8.703665e-02</td>\n",
       "      <td>-5.198924e-01</td>\n",
       "      <td>5.198924e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.047255e+00</td>\n",
       "      <td>1.249922e+00</td>\n",
       "      <td>8.082019e-01</td>\n",
       "      <td>4.538934e-01</td>\n",
       "      <td>9.445668e-02</td>\n",
       "      <td>-1.285939e-01</td>\n",
       "      <td>-2.970230e-01</td>\n",
       "      <td>2.329449e-01</td>\n",
       "      <td>-6.049438e-02</td>\n",
       "      <td>8.155374e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-2.118226e-01</td>\n",
       "      <td>-2.056151e-01</td>\n",
       "      <td>-1.533955e-01</td>\n",
       "      <td>-1.663754e-01</td>\n",
       "      <td>-1.933705e-01</td>\n",
       "      <td>-1.834139e-01</td>\n",
       "      <td>-1.964120e-01</td>\n",
       "      <td>-8.703665e-02</td>\n",
       "      <td>-5.198924e-01</td>\n",
       "      <td>5.198924e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2.667648e+00</td>\n",
       "      <td>1.249922e+00</td>\n",
       "      <td>1.332376e+00</td>\n",
       "      <td>2.446189e+00</td>\n",
       "      <td>3.365297e+01</td>\n",
       "      <td>6.263205e+01</td>\n",
       "      <td>2.235523e+01</td>\n",
       "      <td>3.721131e+01</td>\n",
       "      <td>5.873973e+01</td>\n",
       "      <td>5.842482e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>4.720930e+00</td>\n",
       "      <td>4.863456e+00</td>\n",
       "      <td>6.519097e+00</td>\n",
       "      <td>6.010503e+00</td>\n",
       "      <td>5.171420e+00</td>\n",
       "      <td>5.452149e+00</td>\n",
       "      <td>5.091338e+00</td>\n",
       "      <td>1.148941e+01</td>\n",
       "      <td>1.923475e+00</td>\n",
       "      <td>5.198924e-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 312 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           AGER_TYP    AKT_DAT_KL      ALTER_HH  ALTERSKATEGORIE_FEIN  \\\n",
       "count  7.443050e+05  7.443050e+05  7.443050e+05          7.443050e+05   \n",
       "mean   8.343793e-15 -1.123348e-14 -2.371110e-15          4.843493e-16   \n",
       "std    1.000001e+00  1.000001e+00  1.000001e+00          1.000001e+00   \n",
       "min   -5.731380e-01 -9.423887e-01 -1.419539e+00         -3.087966e+00   \n",
       "25%   -5.731380e-01 -9.423887e-01 -1.419539e+00         -4.315714e-01   \n",
       "50%   -5.731380e-01 -3.943110e-01  2.840275e-01          2.325272e-01   \n",
       "75%    1.047255e+00  1.249922e+00  8.082019e-01          4.538934e-01   \n",
       "max    2.667648e+00  1.249922e+00  1.332376e+00          2.446189e+00   \n",
       "\n",
       "       ANZ_HAUSHALTE_AKTIV  ANZ_HH_TITEL    ANZ_KINDER  ANZ_PERSONEN  \\\n",
       "count         7.443050e+05  7.443050e+05  7.443050e+05  7.443050e+05   \n",
       "mean          5.360674e-15  1.091576e-14  4.510088e-15 -1.590556e-14   \n",
       "std           1.000001e+00  1.000001e+00  1.000001e+00  1.000001e+00   \n",
       "min          -5.435379e-01 -1.285939e-01 -2.970230e-01 -1.486979e+00   \n",
       "25%          -4.159390e-01 -1.285939e-01 -2.970230e-01 -6.270170e-01   \n",
       "50%          -2.883401e-01 -1.285939e-01 -2.970230e-01 -6.270170e-01   \n",
       "75%           9.445668e-02 -1.285939e-01 -2.970230e-01  2.329449e-01   \n",
       "max           3.365297e+01  6.263205e+01  2.235523e+01  3.721131e+01   \n",
       "\n",
       "          ANZ_TITEL        ARBEIT  ...  CAMEO_DEU_2015_8B  CAMEO_DEU_2015_8C  \\\n",
       "count  7.443050e+05  7.443050e+05  ...       7.443050e+05       7.443050e+05   \n",
       "mean  -2.702642e-15 -1.852886e-15  ...      -5.352440e-15       1.058515e-14   \n",
       "std    1.000001e+00  1.000001e+00  ...       1.000001e+00       1.000001e+00   \n",
       "min   -6.049438e-02 -2.200629e+00  ...      -2.118226e-01      -2.056151e-01   \n",
       "25%   -6.049438e-02 -1.898515e-01  ...      -2.118226e-01      -2.056151e-01   \n",
       "50%   -6.049438e-02 -1.898515e-01  ...      -2.118226e-01      -2.056151e-01   \n",
       "75%   -6.049438e-02  8.155374e-01  ...      -2.118226e-01      -2.056151e-01   \n",
       "max    5.873973e+01  5.842482e+00  ...       4.720930e+00       4.863456e+00   \n",
       "\n",
       "       CAMEO_DEU_2015_8D  CAMEO_DEU_2015_9A  CAMEO_DEU_2015_9B  \\\n",
       "count       7.443050e+05       7.443050e+05       7.443050e+05   \n",
       "mean        2.018845e-14       1.703655e-15      -3.483268e-14   \n",
       "std         1.000001e+00       1.000001e+00       1.000001e+00   \n",
       "min        -1.533955e-01      -1.663754e-01      -1.933705e-01   \n",
       "25%        -1.533955e-01      -1.663754e-01      -1.933705e-01   \n",
       "50%        -1.533955e-01      -1.663754e-01      -1.933705e-01   \n",
       "75%        -1.533955e-01      -1.663754e-01      -1.933705e-01   \n",
       "max         6.519097e+00       6.010503e+00       5.171420e+00   \n",
       "\n",
       "       CAMEO_DEU_2015_9C  CAMEO_DEU_2015_9D  CAMEO_DEU_2015_9E  OST_WEST_KZ_O  \\\n",
       "count       7.443050e+05       7.443050e+05       7.443050e+05   7.443050e+05   \n",
       "mean       -3.938322e-15       2.873153e-15       9.391989e-15   2.676556e-14   \n",
       "std         1.000001e+00       1.000001e+00       1.000001e+00   1.000001e+00   \n",
       "min        -1.834139e-01      -1.964120e-01      -8.703665e-02  -5.198924e-01   \n",
       "25%        -1.834139e-01      -1.964120e-01      -8.703665e-02  -5.198924e-01   \n",
       "50%        -1.834139e-01      -1.964120e-01      -8.703665e-02  -5.198924e-01   \n",
       "75%        -1.834139e-01      -1.964120e-01      -8.703665e-02  -5.198924e-01   \n",
       "max         5.452149e+00       5.091338e+00       1.148941e+01   1.923475e+00   \n",
       "\n",
       "       OST_WEST_KZ_W  \n",
       "count   7.443050e+05  \n",
       "mean   -2.676556e-14  \n",
       "std     1.000001e+00  \n",
       "min    -1.923475e+00  \n",
       "25%     5.198924e-01  \n",
       "50%     5.198924e-01  \n",
       "75%     5.198924e-01  \n",
       "max     5.198924e-01  \n",
       "\n",
       "[8 rows x 312 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias_clean.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>ANZ_HH_TITEL</th>\n",
       "      <th>ANZ_KINDER</th>\n",
       "      <th>ANZ_PERSONEN</th>\n",
       "      <th>ANZ_TITEL</th>\n",
       "      <th>ARBEIT</th>\n",
       "      <th>...</th>\n",
       "      <th>CAMEO_DEU_2015_8B</th>\n",
       "      <th>CAMEO_DEU_2015_8C</th>\n",
       "      <th>CAMEO_DEU_2015_8D</th>\n",
       "      <th>CAMEO_DEU_2015_9A</th>\n",
       "      <th>CAMEO_DEU_2015_9B</th>\n",
       "      <th>CAMEO_DEU_2015_9C</th>\n",
       "      <th>CAMEO_DEU_2015_9D</th>\n",
       "      <th>CAMEO_DEU_2015_9E</th>\n",
       "      <th>OST_WEST_KZ_O</th>\n",
       "      <th>OST_WEST_KZ_W</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>...</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "      <td>1.916520e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>-4.238650e-15</td>\n",
       "      <td>3.998211e-15</td>\n",
       "      <td>2.597975e-15</td>\n",
       "      <td>-1.138547e-15</td>\n",
       "      <td>-3.218478e-15</td>\n",
       "      <td>5.157062e-15</td>\n",
       "      <td>3.420783e-15</td>\n",
       "      <td>3.594420e-16</td>\n",
       "      <td>4.665904e-15</td>\n",
       "      <td>3.191686e-16</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.793365e-15</td>\n",
       "      <td>-3.218736e-15</td>\n",
       "      <td>-3.250346e-15</td>\n",
       "      <td>-3.330061e-15</td>\n",
       "      <td>-1.176570e-15</td>\n",
       "      <td>3.265726e-15</td>\n",
       "      <td>5.354518e-17</td>\n",
       "      <td>-4.434903e-15</td>\n",
       "      <td>-2.903817e-15</td>\n",
       "      <td>2.901858e-15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "      <td>1.000003e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-9.660048e-01</td>\n",
       "      <td>-3.250739e-01</td>\n",
       "      <td>-1.174498e+00</td>\n",
       "      <td>-2.897572e+00</td>\n",
       "      <td>-3.164432e-01</td>\n",
       "      <td>-1.052176e-01</td>\n",
       "      <td>-2.383791e-01</td>\n",
       "      <td>-1.812549e+00</td>\n",
       "      <td>-1.162835e-01</td>\n",
       "      <td>-2.144783e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.420366e-01</td>\n",
       "      <td>-1.217859e-01</td>\n",
       "      <td>-1.079096e-01</td>\n",
       "      <td>-7.325602e-02</td>\n",
       "      <td>-6.933771e-02</td>\n",
       "      <td>-7.194872e-02</td>\n",
       "      <td>-1.015705e-01</td>\n",
       "      <td>-8.905522e-02</td>\n",
       "      <td>-2.508160e-01</td>\n",
       "      <td>-3.986986e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-9.660048e-01</td>\n",
       "      <td>-3.250739e-01</td>\n",
       "      <td>-1.174498e+00</td>\n",
       "      <td>-3.513469e-01</td>\n",
       "      <td>-2.359790e-01</td>\n",
       "      <td>-1.052176e-01</td>\n",
       "      <td>-2.383791e-01</td>\n",
       "      <td>-9.896773e-01</td>\n",
       "      <td>-1.162835e-01</td>\n",
       "      <td>-9.984408e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.420366e-01</td>\n",
       "      <td>-1.217859e-01</td>\n",
       "      <td>-1.079096e-01</td>\n",
       "      <td>-7.325602e-02</td>\n",
       "      <td>-6.933771e-02</td>\n",
       "      <td>-7.194872e-02</td>\n",
       "      <td>-1.015705e-01</td>\n",
       "      <td>-8.905522e-02</td>\n",
       "      <td>-2.508160e-01</td>\n",
       "      <td>2.508160e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-2.474429e-01</td>\n",
       "      <td>-3.250739e-01</td>\n",
       "      <td>5.576999e-02</td>\n",
       "      <td>-6.843301e-02</td>\n",
       "      <td>-2.359790e-01</td>\n",
       "      <td>-1.052176e-01</td>\n",
       "      <td>-2.383791e-01</td>\n",
       "      <td>-1.668053e-01</td>\n",
       "      <td>-1.162835e-01</td>\n",
       "      <td>1.479014e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.420366e-01</td>\n",
       "      <td>-1.217859e-01</td>\n",
       "      <td>-1.079096e-01</td>\n",
       "      <td>-7.325602e-02</td>\n",
       "      <td>-6.933771e-02</td>\n",
       "      <td>-7.194872e-02</td>\n",
       "      <td>-1.015705e-01</td>\n",
       "      <td>-8.905522e-02</td>\n",
       "      <td>-2.508160e-01</td>\n",
       "      <td>2.508160e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.189681e+00</td>\n",
       "      <td>-3.250739e-01</td>\n",
       "      <td>8.759486e-01</td>\n",
       "      <td>4.973948e-01</td>\n",
       "      <td>-1.555148e-01</td>\n",
       "      <td>-1.052176e-01</td>\n",
       "      <td>-2.383791e-01</td>\n",
       "      <td>6.560666e-01</td>\n",
       "      <td>-1.162835e-01</td>\n",
       "      <td>1.479014e-01</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.420366e-01</td>\n",
       "      <td>-1.217859e-01</td>\n",
       "      <td>-1.079096e-01</td>\n",
       "      <td>-7.325602e-02</td>\n",
       "      <td>-6.933771e-02</td>\n",
       "      <td>-7.194872e-02</td>\n",
       "      <td>-1.015705e-01</td>\n",
       "      <td>-8.905522e-02</td>\n",
       "      <td>-2.508160e-01</td>\n",
       "      <td>2.508160e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.908243e+00</td>\n",
       "      <td>4.271391e+00</td>\n",
       "      <td>1.696127e+00</td>\n",
       "      <td>4.175275e+00</td>\n",
       "      <td>4.176631e+01</td>\n",
       "      <td>4.919838e+01</td>\n",
       "      <td>1.823360e+01</td>\n",
       "      <td>1.546776e+01</td>\n",
       "      <td>3.755438e+01</td>\n",
       "      <td>7.025955e+00</td>\n",
       "      <td>...</td>\n",
       "      <td>7.040441e+00</td>\n",
       "      <td>8.211133e+00</td>\n",
       "      <td>9.267017e+00</td>\n",
       "      <td>1.365075e+01</td>\n",
       "      <td>1.442217e+01</td>\n",
       "      <td>1.389879e+01</td>\n",
       "      <td>9.845381e+00</td>\n",
       "      <td>1.122899e+01</td>\n",
       "      <td>3.986986e+00</td>\n",
       "      <td>2.508160e-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 312 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           AGER_TYP    AKT_DAT_KL      ALTER_HH  ALTERSKATEGORIE_FEIN  \\\n",
       "count  1.916520e+05  1.916520e+05  1.916520e+05          1.916520e+05   \n",
       "mean  -4.238650e-15  3.998211e-15  2.597975e-15         -1.138547e-15   \n",
       "std    1.000003e+00  1.000003e+00  1.000003e+00          1.000003e+00   \n",
       "min   -9.660048e-01 -3.250739e-01 -1.174498e+00         -2.897572e+00   \n",
       "25%   -9.660048e-01 -3.250739e-01 -1.174498e+00         -3.513469e-01   \n",
       "50%   -2.474429e-01 -3.250739e-01  5.576999e-02         -6.843301e-02   \n",
       "75%    1.189681e+00 -3.250739e-01  8.759486e-01          4.973948e-01   \n",
       "max    1.908243e+00  4.271391e+00  1.696127e+00          4.175275e+00   \n",
       "\n",
       "       ANZ_HAUSHALTE_AKTIV  ANZ_HH_TITEL    ANZ_KINDER  ANZ_PERSONEN  \\\n",
       "count         1.916520e+05  1.916520e+05  1.916520e+05  1.916520e+05   \n",
       "mean         -3.218478e-15  5.157062e-15  3.420783e-15  3.594420e-16   \n",
       "std           1.000003e+00  1.000003e+00  1.000003e+00  1.000003e+00   \n",
       "min          -3.164432e-01 -1.052176e-01 -2.383791e-01 -1.812549e+00   \n",
       "25%          -2.359790e-01 -1.052176e-01 -2.383791e-01 -9.896773e-01   \n",
       "50%          -2.359790e-01 -1.052176e-01 -2.383791e-01 -1.668053e-01   \n",
       "75%          -1.555148e-01 -1.052176e-01 -2.383791e-01  6.560666e-01   \n",
       "max           4.176631e+01  4.919838e+01  1.823360e+01  1.546776e+01   \n",
       "\n",
       "          ANZ_TITEL        ARBEIT  ...  CAMEO_DEU_2015_8B  CAMEO_DEU_2015_8C  \\\n",
       "count  1.916520e+05  1.916520e+05  ...       1.916520e+05       1.916520e+05   \n",
       "mean   4.665904e-15  3.191686e-16  ...      -1.793365e-15      -3.218736e-15   \n",
       "std    1.000003e+00  1.000003e+00  ...       1.000003e+00       1.000003e+00   \n",
       "min   -1.162835e-01 -2.144783e+00  ...      -1.420366e-01      -1.217859e-01   \n",
       "25%   -1.162835e-01 -9.984408e-01  ...      -1.420366e-01      -1.217859e-01   \n",
       "50%   -1.162835e-01  1.479014e-01  ...      -1.420366e-01      -1.217859e-01   \n",
       "75%   -1.162835e-01  1.479014e-01  ...      -1.420366e-01      -1.217859e-01   \n",
       "max    3.755438e+01  7.025955e+00  ...       7.040441e+00       8.211133e+00   \n",
       "\n",
       "       CAMEO_DEU_2015_8D  CAMEO_DEU_2015_9A  CAMEO_DEU_2015_9B  \\\n",
       "count       1.916520e+05       1.916520e+05       1.916520e+05   \n",
       "mean       -3.250346e-15      -3.330061e-15      -1.176570e-15   \n",
       "std         1.000003e+00       1.000003e+00       1.000003e+00   \n",
       "min        -1.079096e-01      -7.325602e-02      -6.933771e-02   \n",
       "25%        -1.079096e-01      -7.325602e-02      -6.933771e-02   \n",
       "50%        -1.079096e-01      -7.325602e-02      -6.933771e-02   \n",
       "75%        -1.079096e-01      -7.325602e-02      -6.933771e-02   \n",
       "max         9.267017e+00       1.365075e+01       1.442217e+01   \n",
       "\n",
       "       CAMEO_DEU_2015_9C  CAMEO_DEU_2015_9D  CAMEO_DEU_2015_9E  OST_WEST_KZ_O  \\\n",
       "count       1.916520e+05       1.916520e+05       1.916520e+05   1.916520e+05   \n",
       "mean        3.265726e-15       5.354518e-17      -4.434903e-15  -2.903817e-15   \n",
       "std         1.000003e+00       1.000003e+00       1.000003e+00   1.000003e+00   \n",
       "min        -7.194872e-02      -1.015705e-01      -8.905522e-02  -2.508160e-01   \n",
       "25%        -7.194872e-02      -1.015705e-01      -8.905522e-02  -2.508160e-01   \n",
       "50%        -7.194872e-02      -1.015705e-01      -8.905522e-02  -2.508160e-01   \n",
       "75%        -7.194872e-02      -1.015705e-01      -8.905522e-02  -2.508160e-01   \n",
       "max         1.389879e+01       9.845381e+00       1.122899e+01   3.986986e+00   \n",
       "\n",
       "       OST_WEST_KZ_W  \n",
       "count   1.916520e+05  \n",
       "mean    2.901858e-15  \n",
       "std     1.000003e+00  \n",
       "min    -3.986986e+00  \n",
       "25%     2.508160e-01  \n",
       "50%     2.508160e-01  \n",
       "75%     2.508160e-01  \n",
       "max     2.508160e-01  \n",
       "\n",
       "[8 rows x 312 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers_clean.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save cleaned data\n",
    "azdias_clean.to_csv('azdias_cleaned.csv', index=False)\n",
    "customers_clean.to_csv('customers_cleaned.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 1: Customer Segmentation Report\n",
    "\n",
    "The main bulk of your analysis will come in this part of the project. Here, you should use unsupervised learning techniques to describe the relationship between the demographics of the company's existing customers and the general population of Germany. By the end of this part, you should be able to describe parts of the general population that are more likely to be part of the mail-order company's main customer base, and which parts of the general population are less so."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the data\n",
    "azdias = pd.read_csv('azdias_cleaned.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>ANZ_HH_TITEL</th>\n",
       "      <th>ANZ_KINDER</th>\n",
       "      <th>ANZ_PERSONEN</th>\n",
       "      <th>ANZ_TITEL</th>\n",
       "      <th>ARBEIT</th>\n",
       "      <th>...</th>\n",
       "      <th>CAMEO_DEU_2015_8B</th>\n",
       "      <th>CAMEO_DEU_2015_8C</th>\n",
       "      <th>CAMEO_DEU_2015_8D</th>\n",
       "      <th>CAMEO_DEU_2015_9A</th>\n",
       "      <th>CAMEO_DEU_2015_9B</th>\n",
       "      <th>CAMEO_DEU_2015_9C</th>\n",
       "      <th>CAMEO_DEU_2015_9D</th>\n",
       "      <th>CAMEO_DEU_2015_9E</th>\n",
       "      <th>OST_WEST_KZ_O</th>\n",
       "      <th>OST_WEST_KZ_W</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.573138</td>\n",
       "      <td>1.249922</td>\n",
       "      <td>-1.419539</td>\n",
       "      <td>1.560724</td>\n",
       "      <td>0.158256</td>\n",
       "      <td>-0.128594</td>\n",
       "      <td>-0.297023</td>\n",
       "      <td>0.232945</td>\n",
       "      <td>-0.060494</td>\n",
       "      <td>-0.189851</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.211823</td>\n",
       "      <td>-0.205615</td>\n",
       "      <td>-0.153395</td>\n",
       "      <td>-0.166375</td>\n",
       "      <td>-0.19337</td>\n",
       "      <td>-0.183414</td>\n",
       "      <td>-0.196412</td>\n",
       "      <td>-0.087037</td>\n",
       "      <td>-0.519892</td>\n",
       "      <td>0.519892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.573138</td>\n",
       "      <td>1.249922</td>\n",
       "      <td>0.808202</td>\n",
       "      <td>0.675260</td>\n",
       "      <td>0.094457</td>\n",
       "      <td>-0.128594</td>\n",
       "      <td>-0.297023</td>\n",
       "      <td>-0.627017</td>\n",
       "      <td>-0.060494</td>\n",
       "      <td>-0.189851</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.211823</td>\n",
       "      <td>-0.205615</td>\n",
       "      <td>-0.153395</td>\n",
       "      <td>-0.166375</td>\n",
       "      <td>-0.19337</td>\n",
       "      <td>-0.183414</td>\n",
       "      <td>-0.196412</td>\n",
       "      <td>-0.087037</td>\n",
       "      <td>-0.519892</td>\n",
       "      <td>0.519892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.857451</td>\n",
       "      <td>-0.942389</td>\n",
       "      <td>0.284028</td>\n",
       "      <td>-0.210205</td>\n",
       "      <td>-0.479738</td>\n",
       "      <td>-0.128594</td>\n",
       "      <td>-0.297023</td>\n",
       "      <td>-1.486979</td>\n",
       "      <td>-0.060494</td>\n",
       "      <td>-1.195240</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.211823</td>\n",
       "      <td>-0.205615</td>\n",
       "      <td>-0.153395</td>\n",
       "      <td>-0.166375</td>\n",
       "      <td>-0.19337</td>\n",
       "      <td>-0.183414</td>\n",
       "      <td>-0.196412</td>\n",
       "      <td>-0.087037</td>\n",
       "      <td>-0.519892</td>\n",
       "      <td>0.519892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.573138</td>\n",
       "      <td>-0.942389</td>\n",
       "      <td>1.201333</td>\n",
       "      <td>0.011161</td>\n",
       "      <td>-0.352140</td>\n",
       "      <td>-0.128594</td>\n",
       "      <td>-0.297023</td>\n",
       "      <td>1.952869</td>\n",
       "      <td>-0.060494</td>\n",
       "      <td>0.815537</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.211823</td>\n",
       "      <td>-0.205615</td>\n",
       "      <td>-0.153395</td>\n",
       "      <td>-0.166375</td>\n",
       "      <td>-0.19337</td>\n",
       "      <td>-0.183414</td>\n",
       "      <td>-0.196412</td>\n",
       "      <td>-0.087037</td>\n",
       "      <td>-0.519892</td>\n",
       "      <td>0.519892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.667648</td>\n",
       "      <td>-0.942389</td>\n",
       "      <td>-0.109103</td>\n",
       "      <td>-0.874304</td>\n",
       "      <td>-0.224541</td>\n",
       "      <td>-0.128594</td>\n",
       "      <td>-0.297023</td>\n",
       "      <td>-0.627017</td>\n",
       "      <td>-0.060494</td>\n",
       "      <td>-1.195240</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.211823</td>\n",
       "      <td>4.863456</td>\n",
       "      <td>-0.153395</td>\n",
       "      <td>-0.166375</td>\n",
       "      <td>-0.19337</td>\n",
       "      <td>-0.183414</td>\n",
       "      <td>-0.196412</td>\n",
       "      <td>-0.087037</td>\n",
       "      <td>-0.519892</td>\n",
       "      <td>0.519892</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 312 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTERSKATEGORIE_FEIN  ANZ_HAUSHALTE_AKTIV  \\\n",
       "0 -0.573138    1.249922 -1.419539              1.560724             0.158256   \n",
       "1 -0.573138    1.249922  0.808202              0.675260             0.094457   \n",
       "2  1.857451   -0.942389  0.284028             -0.210205            -0.479738   \n",
       "3 -0.573138   -0.942389  1.201333              0.011161            -0.352140   \n",
       "4  2.667648   -0.942389 -0.109103             -0.874304            -0.224541   \n",
       "\n",
       "   ANZ_HH_TITEL  ANZ_KINDER  ANZ_PERSONEN  ANZ_TITEL    ARBEIT  ...  \\\n",
       "0     -0.128594   -0.297023      0.232945  -0.060494 -0.189851  ...   \n",
       "1     -0.128594   -0.297023     -0.627017  -0.060494 -0.189851  ...   \n",
       "2     -0.128594   -0.297023     -1.486979  -0.060494 -1.195240  ...   \n",
       "3     -0.128594   -0.297023      1.952869  -0.060494  0.815537  ...   \n",
       "4     -0.128594   -0.297023     -0.627017  -0.060494 -1.195240  ...   \n",
       "\n",
       "   CAMEO_DEU_2015_8B  CAMEO_DEU_2015_8C  CAMEO_DEU_2015_8D  CAMEO_DEU_2015_9A  \\\n",
       "0          -0.211823          -0.205615          -0.153395          -0.166375   \n",
       "1          -0.211823          -0.205615          -0.153395          -0.166375   \n",
       "2          -0.211823          -0.205615          -0.153395          -0.166375   \n",
       "3          -0.211823          -0.205615          -0.153395          -0.166375   \n",
       "4          -0.211823           4.863456          -0.153395          -0.166375   \n",
       "\n",
       "   CAMEO_DEU_2015_9B  CAMEO_DEU_2015_9C  CAMEO_DEU_2015_9D  CAMEO_DEU_2015_9E  \\\n",
       "0           -0.19337          -0.183414          -0.196412          -0.087037   \n",
       "1           -0.19337          -0.183414          -0.196412          -0.087037   \n",
       "2           -0.19337          -0.183414          -0.196412          -0.087037   \n",
       "3           -0.19337          -0.183414          -0.196412          -0.087037   \n",
       "4           -0.19337          -0.183414          -0.196412          -0.087037   \n",
       "\n",
       "   OST_WEST_KZ_O  OST_WEST_KZ_W  \n",
       "0      -0.519892       0.519892  \n",
       "1      -0.519892       0.519892  \n",
       "2      -0.519892       0.519892  \n",
       "3      -0.519892       0.519892  \n",
       "4      -0.519892       0.519892  \n",
       "\n",
       "[5 rows x 312 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "azdias.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "customers = pd.read_csv('customers_cleaned.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>ANZ_HH_TITEL</th>\n",
       "      <th>ANZ_KINDER</th>\n",
       "      <th>ANZ_PERSONEN</th>\n",
       "      <th>ANZ_TITEL</th>\n",
       "      <th>ARBEIT</th>\n",
       "      <th>...</th>\n",
       "      <th>CAMEO_DEU_2015_8B</th>\n",
       "      <th>CAMEO_DEU_2015_8C</th>\n",
       "      <th>CAMEO_DEU_2015_8D</th>\n",
       "      <th>CAMEO_DEU_2015_9A</th>\n",
       "      <th>CAMEO_DEU_2015_9B</th>\n",
       "      <th>CAMEO_DEU_2015_9C</th>\n",
       "      <th>CAMEO_DEU_2015_9D</th>\n",
       "      <th>CAMEO_DEU_2015_9E</th>\n",
       "      <th>OST_WEST_KZ_O</th>\n",
       "      <th>OST_WEST_KZ_W</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.189681</td>\n",
       "      <td>-0.325074</td>\n",
       "      <td>0.192466</td>\n",
       "      <td>-0.068433</td>\n",
       "      <td>-0.235979</td>\n",
       "      <td>-0.105218</td>\n",
       "      <td>-0.238379</td>\n",
       "      <td>-0.166805</td>\n",
       "      <td>-0.116283</td>\n",
       "      <td>-2.144783</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.142037</td>\n",
       "      <td>-0.121786</td>\n",
       "      <td>-0.10791</td>\n",
       "      <td>-0.073256</td>\n",
       "      <td>-0.069338</td>\n",
       "      <td>-0.071949</td>\n",
       "      <td>-0.10157</td>\n",
       "      <td>-0.089055</td>\n",
       "      <td>-0.250816</td>\n",
       "      <td>0.250816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.966005</td>\n",
       "      <td>4.271391</td>\n",
       "      <td>0.329163</td>\n",
       "      <td>-0.068433</td>\n",
       "      <td>-0.235979</td>\n",
       "      <td>-0.105218</td>\n",
       "      <td>-0.238379</td>\n",
       "      <td>0.656067</td>\n",
       "      <td>-0.116283</td>\n",
       "      <td>0.147901</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.142037</td>\n",
       "      <td>-0.121786</td>\n",
       "      <td>-0.10791</td>\n",
       "      <td>-0.073256</td>\n",
       "      <td>-0.069338</td>\n",
       "      <td>-0.071949</td>\n",
       "      <td>-0.10157</td>\n",
       "      <td>-0.089055</td>\n",
       "      <td>-0.250816</td>\n",
       "      <td>0.250816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.966005</td>\n",
       "      <td>-0.325074</td>\n",
       "      <td>-0.354319</td>\n",
       "      <td>-2.897572</td>\n",
       "      <td>-0.235979</td>\n",
       "      <td>-0.105218</td>\n",
       "      <td>-0.238379</td>\n",
       "      <td>-0.989677</td>\n",
       "      <td>-0.116283</td>\n",
       "      <td>0.147901</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.142037</td>\n",
       "      <td>-0.121786</td>\n",
       "      <td>-0.10791</td>\n",
       "      <td>-0.073256</td>\n",
       "      <td>-0.069338</td>\n",
       "      <td>-0.071949</td>\n",
       "      <td>-0.10157</td>\n",
       "      <td>-0.089055</td>\n",
       "      <td>-0.250816</td>\n",
       "      <td>0.250816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.471119</td>\n",
       "      <td>-0.325074</td>\n",
       "      <td>-0.080926</td>\n",
       "      <td>-0.634261</td>\n",
       "      <td>-0.316443</td>\n",
       "      <td>-0.105218</td>\n",
       "      <td>-0.238379</td>\n",
       "      <td>-1.812549</td>\n",
       "      <td>-0.116283</td>\n",
       "      <td>-2.144783</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.142037</td>\n",
       "      <td>-0.121786</td>\n",
       "      <td>-0.10791</td>\n",
       "      <td>-0.073256</td>\n",
       "      <td>-0.069338</td>\n",
       "      <td>-0.071949</td>\n",
       "      <td>-0.10157</td>\n",
       "      <td>-0.089055</td>\n",
       "      <td>-0.250816</td>\n",
       "      <td>0.250816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.966005</td>\n",
       "      <td>-0.325074</td>\n",
       "      <td>1.559431</td>\n",
       "      <td>1.063223</td>\n",
       "      <td>0.246806</td>\n",
       "      <td>-0.105218</td>\n",
       "      <td>-0.238379</td>\n",
       "      <td>1.478939</td>\n",
       "      <td>-0.116283</td>\n",
       "      <td>0.147901</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.142037</td>\n",
       "      <td>-0.121786</td>\n",
       "      <td>-0.10791</td>\n",
       "      <td>-0.073256</td>\n",
       "      <td>-0.069338</td>\n",
       "      <td>-0.071949</td>\n",
       "      <td>-0.10157</td>\n",
       "      <td>-0.089055</td>\n",
       "      <td>-0.250816</td>\n",
       "      <td>0.250816</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 312 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTERSKATEGORIE_FEIN  ANZ_HAUSHALTE_AKTIV  \\\n",
       "0  1.189681   -0.325074  0.192466             -0.068433            -0.235979   \n",
       "1 -0.966005    4.271391  0.329163             -0.068433            -0.235979   \n",
       "2 -0.966005   -0.325074 -0.354319             -2.897572            -0.235979   \n",
       "3  0.471119   -0.325074 -0.080926             -0.634261            -0.316443   \n",
       "4 -0.966005   -0.325074  1.559431              1.063223             0.246806   \n",
       "\n",
       "   ANZ_HH_TITEL  ANZ_KINDER  ANZ_PERSONEN  ANZ_TITEL    ARBEIT  ...  \\\n",
       "0     -0.105218   -0.238379     -0.166805  -0.116283 -2.144783  ...   \n",
       "1     -0.105218   -0.238379      0.656067  -0.116283  0.147901  ...   \n",
       "2     -0.105218   -0.238379     -0.989677  -0.116283  0.147901  ...   \n",
       "3     -0.105218   -0.238379     -1.812549  -0.116283 -2.144783  ...   \n",
       "4     -0.105218   -0.238379      1.478939  -0.116283  0.147901  ...   \n",
       "\n",
       "   CAMEO_DEU_2015_8B  CAMEO_DEU_2015_8C  CAMEO_DEU_2015_8D  CAMEO_DEU_2015_9A  \\\n",
       "0          -0.142037          -0.121786           -0.10791          -0.073256   \n",
       "1          -0.142037          -0.121786           -0.10791          -0.073256   \n",
       "2          -0.142037          -0.121786           -0.10791          -0.073256   \n",
       "3          -0.142037          -0.121786           -0.10791          -0.073256   \n",
       "4          -0.142037          -0.121786           -0.10791          -0.073256   \n",
       "\n",
       "   CAMEO_DEU_2015_9B  CAMEO_DEU_2015_9C  CAMEO_DEU_2015_9D  CAMEO_DEU_2015_9E  \\\n",
       "0          -0.069338          -0.071949           -0.10157          -0.089055   \n",
       "1          -0.069338          -0.071949           -0.10157          -0.089055   \n",
       "2          -0.069338          -0.071949           -0.10157          -0.089055   \n",
       "3          -0.069338          -0.071949           -0.10157          -0.089055   \n",
       "4          -0.069338          -0.071949           -0.10157          -0.089055   \n",
       "\n",
       "   OST_WEST_KZ_O  OST_WEST_KZ_W  \n",
       "0      -0.250816       0.250816  \n",
       "1      -0.250816       0.250816  \n",
       "2      -0.250816       0.250816  \n",
       "3      -0.250816       0.250816  \n",
       "4      -0.250816       0.250816  \n",
       "\n",
       "[5 rows x 312 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(744305, 312)\n",
      "(191652, 312)\n"
     ]
    }
   ],
   "source": [
    "print(azdias.shape)\n",
    "print(customers.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dimensionality Reduction - PCA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data is now ready for dimensionality reduction. \n",
    "Use sklearn's PCA class to apply principal component analysis on the data, thus finding the vectors of maximal variance in the data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Perform PCA \n",
    "pca = PCA()\n",
    "azdias_pca = pca.fit_transform(azdias)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using this function from Udacity's class\n",
    "def scree_plot(pca, num_components=None):\n",
    "    '''\n",
    "    Creates a scree plot associated with the principal components \n",
    "    \n",
    "    INPUT: pca - the result of instantian of PCA in scikit learn\n",
    "            \n",
    "    OUTPUT:\n",
    "            None\n",
    "    '''\n",
    "    if (num_components == None):\n",
    "        num_components = len(pca.explained_variance_ratio_)\n",
    "    ind = np.arange(num_components)\n",
    "    vals = pca.explained_variance_ratio_[:num_components]\n",
    " \n",
    "    plt.figure(figsize=(18, 12))\n",
    "    ax = plt.subplot(111)\n",
    "    cumvals = np.cumsum(vals)\n",
    "    ax.bar(ind, vals)\n",
    "    ax.plot(ind, cumvals)\n",
    " \n",
    "    ax.set_xlabel(\"Principal Component\")\n",
    "    ax.set_ylabel(\"Variance Explained (%)\")\n",
    "    plt.title('Explained Variance Per Principal Component')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "scree_plot(pca)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10 components explains 26.902190737478104% of variance\n",
      "30 components explains 41.69724273591772% of variance\n",
      "50 components explains 50.50828836219898% of variance\n",
      "70 components explains 57.43200162028395% of variance\n",
      "90 components explains 63.942451103513% of variance\n",
      "110 components explains 70.1637418158014% of variance\n",
      "130 components explains 75.7430885215517% of variance\n",
      "150 components explains 80.72916389575636% of variance\n",
      "170 components explains 85.17967055399541% of variance\n",
      "190 components explains 89.07342290572663% of variance\n",
      "210 components explains 92.3025479234256% of variance\n",
      "230 components explains 94.80731100329045% of variance\n",
      "250 components explains 96.75943904448626% of variance\n",
      "270 components explains 98.25973590670007% of variance\n",
      "290 components explains 99.35660537025502% of variance\n",
      "310 components explains 100.0% of variance\n"
     ]
    }
   ],
   "source": [
    "# Variance Explained by increasing components\n",
    "for i in np.arange(10, len(pca.explained_variance_ratio_), 20):\n",
    "    print('{} components explains {}% of variance'.format(i,(pca.explained_variance_ratio_[:i].sum()) * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8721086702452737"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Choosing number of componetns\n",
    "pca.explained_variance_ratio_[:180].sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Choosing 180 components, the variance explained is around 87%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Re-apply PCA to the data while selecting for number of components to retain.\n",
    "pca = PCA(n_components=180)\n",
    "azdias_pca = pca.fit_transform(azdias)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x864 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "scree_plot(pca,num_components=180)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With 180 componets we were able to explained 87% of variability, we have reduced the number of features from 312 to only 180\n",
    "- We have reduced the data by 132 features and still kept 87 % of variance explained.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Visualize PCA Results\n",
    "def pca_results(df, pca, component_no, features_no):\n",
    "    '''\n",
    "    Create a DataFrame of the PCA results\n",
    "    Visualizes the PCA results for a speacific component number for a speacific number of features\n",
    "    \n",
    "    '''\n",
    "    \n",
    "    pca_comp = pd.DataFrame(np.round(pca.components_, 4), columns=df.keys()).iloc[component_no - 1]\n",
    "    pca_comp.sort_values(ascending=False, inplace=True)\n",
    "    # Capturing most interesting features which at the end and beginning of sorted list\n",
    "    pca_comp = pd.concat([pca_comp.head(features_no), pca_comp.tail(features_no)])\n",
    "    \n",
    "    pca_comp.plot(kind='bar', \n",
    "                  title='Most {} weighted features for PCA component {}'.format(features_no*2, component_no),\n",
    "                  figsize=(12, 6));\n",
    "    plt.show()\n",
    "    \n",
    "    return pca_comp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "HH_EINKOMMEN_SCORE    0.1610\n",
       "CAMEO_DEUG_2015       0.1595\n",
       "KBA13_ANTG4           0.1430\n",
       "FINANZ_HAUSBAUER      0.1244\n",
       "ARBEIT                0.1163\n",
       "KBA05_GBZ            -0.1489\n",
       "KBA13_ANTG1          -0.1509\n",
       "LP_STATUS_FEIN       -0.1539\n",
       "KBA05_AUTOQUOT       -0.1557\n",
       "KBA05_ANTG1          -0.1558\n",
       "Name: 0, dtype: float64"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Map weights for the first principal component to corresponding feature names\n",
    "# and then print the linked values, sorted by weight.\n",
    "pca_results(azdias, pca, 1, 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Map weights for the first principal component to corresponding feature names\n",
    "# and then print the linked values, sorted by weight.\n",
    "# HINT: Try defining a function here or in a new cell that you can reuse in the\n",
    "# other cells.\n",
    "\n",
    "# Map weights for the third principal component to corresponding feature names\n",
    "# and then print the linked values, sorted by weight.\n",
    "\n",
    "# Plot features_no features with the most positive and negative variance\n",
    "# for a component_no PCA component\n",
    "def pca_results(df, pca, component_no, features_no, show_plot=True):\n",
    "    \n",
    "    # Build a dataframe with features_no features capturing most variability\n",
    "    # for a given component number (component_no)\n",
    "    pca_comp = pd.DataFrame(np.round(pca.components_, 4), columns=df.keys()).iloc[component_no - 1]\n",
    "    pca_comp.sort_values(ascending=False, inplace=True)\n",
    "    pca_comp = pd.concat([pca_comp.head(features_no), pca_comp.tail(features_no)])\n",
    "    \n",
    "    if show_plot:\n",
    "        # Plot the result\n",
    "        pca_comp.plot(kind='bar', \n",
    "                  title='Most {} weighted features for PCA component {}'.format(features_no*2, component_no),\n",
    "                  figsize=(12, 6))\n",
    "        plt.show()\n",
    "    \n",
    "    return pca_comp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KBA13_CCM_1200      0.2470\n",
      "KBA13_CCM_2500      0.2227\n",
      "RETOURTYP_BK_S      0.2086\n",
      "GREEN_AVANTGARDE    0.2043\n",
      "KBA13_KW_60         0.1744\n",
      "KBA13_CCM_1000     -0.1530\n",
      "KBA13_AUDI         -0.1558\n",
      "KBA13_KW_120       -0.1997\n",
      "UMFELD_ALT         -0.2972\n",
      "KBA13_KW_50        -0.3114\n",
      "Name: 179, dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HH_EINKOMMEN_SCORE    0.1610\n",
      "CAMEO_DEUG_2015       0.1595\n",
      "KBA13_ANTG4           0.1430\n",
      "FINANZ_HAUSBAUER      0.1244\n",
      "ARBEIT                0.1163\n",
      "KBA05_GBZ            -0.1489\n",
      "KBA13_ANTG1          -0.1509\n",
      "LP_STATUS_FEIN       -0.1539\n",
      "KBA05_AUTOQUOT       -0.1557\n",
      "KBA05_ANTG1          -0.1558\n",
      "Name: 0, dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KBA13_HERST_BMW_BENZ           0.2009\n",
      "KBA13_SEG_OBEREMITTELKLASSE    0.1758\n",
      "KBA13_BMW                      0.1745\n",
      "KBA13_SITZE_4                  0.1592\n",
      "KBA13_SEG_SPORTWAGEN           0.1559\n",
      "KBA13_HERST_FORD_OPEL         -0.1196\n",
      "KBA13_ALTERHALTER_60          -0.1224\n",
      "KBA13_HALTER_20               -0.1312\n",
      "KBA13_KMH_180                 -0.1340\n",
      "KBA13_KMH_140_210             -0.1519\n",
      "Name: 1, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Check first 3 components\n",
    "for i in np.arange(0, 3, 1):\n",
    "    res = pca_results(azdias, pca, i, 5)\n",
    "    print(res)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Interprete PCA Results\n",
    "\n",
    "The top 10 weighted features for the PCA for the first 3 components."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### COMPONENT 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***Positively correlated features***\n",
    "- **_KBA13_CCM_1200_** -  share of cars with 1000ccm to 1199ccm within the PLZ8\n",
    "- **_KBA13_CCM_2500_** - share of cars with 2000ccm to 2499ccm within the PLZ8\n",
    "- **_RETOURTYP_BK_S_** - return type- how likely a shopper is going to return to shop again \n",
    "- **_GREEN_AVANTGARDE_** - the environmental sustainability is the dominating movement in the youth of these consumers\n",
    "- **_KBA13_KW_60_** - share of cars with an engine power between 51 and 60 KW - PLZ8\n",
    "\n",
    "***Negatively correlated features***\n",
    "- **_KBA13_CCM_1000_** - share of cars with less than 1000ccm within the PLZ8\n",
    "- **_KBA13_AUDI_** - share of AUDI within the PLZ8\n",
    "- **_KBA13_KW_120_** - share of cars with an engine power between 111 and 120 KW - PLZ8\n",
    "- **_UMFELD_ALT_** -\n",
    "- **_KBA13_KW_50_** - share of cars with an engine power between 41 and 50 KW - PLZ8\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### COMPONENT 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***Positively correlated features***\n",
    "- **_HH_EINKOMMEN_SCORE_** - estimated household net income\n",
    "- **_CAMEO_DEUG_2015_** - Typology clasification based on classes (upper/lower class)\n",
    "- **_KBA13_ANTG4_** - unknown attributes with associated with car in PLZ8\n",
    "- **_FINANZ_HAUSBAUER_** - financial typology: main focus is the own house\n",
    "- **_ARBEIT_** - share of unemployed person in the community\t\n",
    "\n",
    "***Negatively correlated features***\n",
    "- **_KBA05_GBZ_** - number of buildings in the microcell\t\n",
    "- **_KBA13_ANTG1_** - share of unknown car in PLZ8\n",
    "- **_LP_STATUS_FEIN_** - social status \n",
    "- **_KBA05_AUTOQUOT_** - share of cars per household\t\n",
    "- **_KBA05_ANTG1_50_** - number of 1-2 family houses in the cell\t"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### COMPONENT 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***Positively correlated features***\n",
    "- **_KBA13_HERST_BMW_BENZ_** - share of BMW & Mercedes Benz within the PLZ8\n",
    "- **_KBA13_SEG_OBEREMITTELKLASSE_** - share of upper middle class cars and upper class cars (BMW5er, BMW7er etc.)\n",
    "- **_KBA13_BMW_** - share of BMW within the PLZ8\t\n",
    "- **_KBA13_SITZE_4_** - number of cars with less than 5 seats in the PLZ8\n",
    "- **_KBA13_SEG_SPORTWAGEN_** - share of sportscars within the PLZ8\t\n",
    "\n",
    "***Negatively correlated features***\n",
    "- **_KBA13_HERST_FORD_OPEL_** - share of Ford & Opel/Vauxhall within the PLZ8\t\t\n",
    "- **_KBA13_ALTERHALTER_60_** - share of car owners between 46 and 60 within the PLZ8\n",
    "- **_KBA13_HALTER_20_** - share of car owners below 21 within the PLZ8\n",
    "- **_KBA13_KMH_180_** - share of cars with max speed between 110 km/h and 180km/h within the PLZ8\t\n",
    "- **_KBA13_KMH_140_210_** - share of cars with max speed between 140 and 210 km/h within the PLZ8"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Clustering"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apply k-means clustering to the dataset and use the average within-cluster distances from each point to their assigned cluster's centroid to decide on a number of clusters to keep."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### K-Means - General population (azdias)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [],
   "source": [
    "# K-Means fucntion from Udacity k-means\n",
    "def get_kmeans_score(data, center):\n",
    "    '''\n",
    "    returns the kmeans score regarding SSE for points to centers\n",
    "    INPUT:\n",
    "        data - the dataset you want to fit kmeans to\n",
    "        center - the number of centers you want (the k value)\n",
    "    OUTPUT:\n",
    "        score - the SSE score for the kmeans model fit to the data\n",
    "    '''\n",
    "    #instantiate kmeans\n",
    "    kmeans = KMeans(n_clusters=center, max_iter=30) \n",
    "\n",
    "    # Then fit the model to your data using the fit method\n",
    "    model = kmeans.fit(data)\n",
    "    \n",
    "    # Obtain a score related to the model fit\n",
    "    score = np.abs(model.score(data))\n",
    "    \n",
    "    return score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Over a number of different cluster counts...\n",
    "# run k-means clustering on the data and...\n",
    "# compute the average within-cluster distances.\n",
    "\n",
    "scores = []\n",
    "centers = np.arange(1,30,2)\n",
    "\n",
    "for center in centers:\n",
    "    scores.append(get_kmeans_score(azdias_pca, center))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x396 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Investigate the change in within-cluster distance across number of clusters.\n",
    "plt.plot(centers, scores, linestyle='--', marker='o', color='b');\n",
    "plt.xlabel('Centers');\n",
    "plt.ylabel('SSE');\n",
    "plt.title('SSE vs. Centers');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To find the right number of clusters, Elbow Method were used. Unfortunatelly there is no clearly distinguishable elbow in the plot, nevertheless the plot starts to flatten more after K = 22 so 22 is chosen as the number of clusters for the full dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-3.20821035, -3.60203422, -5.14383919, ...,  0.72651149,\n",
       "        -0.73178889, -0.75640715],\n",
       "       [-3.05669414, -0.97492101,  1.38158378, ..., -0.14130874,\n",
       "         0.21353974, -0.85775866],\n",
       "       [-3.01630546, -5.1754569 ,  2.57128741, ..., -0.83877368,\n",
       "         0.35747054,  0.2872013 ],\n",
       "       ...,\n",
       "       [-3.94199963,  2.36402336, -0.79743758, ..., -0.69291247,\n",
       "        -0.87816196, -1.02325298],\n",
       "       [ 5.45383941,  2.35557783,  2.25302965, ..., -1.23103393,\n",
       "         0.46970119, -0.98848325],\n",
       "       [-6.14226593,  0.73065701, -6.88825624, ...,  2.29241563,\n",
       "        -0.11851782, -0.63637582]])"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a toy dataset to reduce computation time\n",
    "# Reduce the size to 25%\n",
    "azdias_pca_sample = azdias_pca[np.random.choice(azdias_pca.shape[0], int(azdias_pca.shape[0]*0.25), replace=False)]\n",
    "azdias_pca_sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 10 k, with SSE 42337008.71951024\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 11 k, with SSE 42108385.569391504\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 12 k, with SSE 41906210.49174486\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 13 k, with SSE 41654754.68684188\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 14 k, with SSE 41508817.79501742\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 15 k, with SSE 41285869.88956661\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 16 k, with SSE 41241457.3415008\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 17 k, with SSE 40966761.46979014\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 18 k, with SSE 40845352.977612875\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 19 k, with SSE 40718156.47682739\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 20 k, with SSE 40577453.279080756\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 21 k, with SSE 40513453.43081363\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 22 k, with SSE 40394317.27126055\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 23 k, with SSE 40303808.14636363\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 24 k, with SSE 40113295.162932016\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 25 k, with SSE 40007746.871411614\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 26 k, with SSE 39924864.91257075\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 27 k, with SSE 39764339.511834346\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 28 k, with SSE 39704357.21721037\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 29 k, with SSE 39630365.446084544\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Clustering done for 30 k, with SSE 39517074.27236323\n"
     ]
    }
   ],
   "source": [
    "# Over a number of different cluster counts...\n",
    "    # run k-means clustering on the data and...\n",
    "    # compute the average within-cluster distances.\n",
    "\n",
    "sse = [] # Sum of Squared Errors\n",
    "k_range = np.arange(10, 31)\n",
    "\n",
    "for k in k_range:\n",
    "    kmeans = KMeans(k, random_state=1234, max_iter=30, n_jobs=-1).fit(azdias_pca_sample)\n",
    "    score = np.abs(kmeans.score(azdias_pca_sample))\n",
    "    sse.append(score)\n",
    "    print('Clustering done for {} k, with SSE {}'.format(k, score))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Investigate the change in within-cluster distance across number of clusters.\n",
    "# Use the Elbow Method to identify the right value of k\n",
    "plt.figure(figsize=(18, 8))\n",
    "plt.xticks(np.arange(0, k_range[-1]+1, step=1))\n",
    "plt.xlabel('K')\n",
    "plt.ylabel('SSE')\n",
    "plt.title('SSE vs. K')\n",
    "plt.plot(k_range, sse, linestyle='-', marker='o');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By using elbow method we can see a strong leveling off at 24 clusters, which suggests 24 clusters should be used"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    }
   ],
   "source": [
    "# Re-fit the k-means model with the selected number of clusters and obtain\n",
    "k = 24\n",
    "kmeans = KMeans(k, random_state=1234, max_iter=30, n_jobs=-1).fit(azdias_pca)\n",
    "population_clusters = kmeans.predict(azdias_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Apply all the process to the customers dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the cleaned customer dataset\n",
    "customers = pd.read_csv('customers_cleaned.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Applying PCA\n",
    "customers_pca = pca.transform(customers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:973: FutureWarning: 'n_jobs' was deprecated in version 0.23 and will be removed in 0.25.\n",
      "  warnings.warn(\"'n_jobs' was deprecated in version 0.23 and will be\"\n"
     ]
    }
   ],
   "source": [
    "### Step 3: Clustering\n",
    "# Cluster predictions for the general population demographics data.\n",
    "kmeans = KMeans(k, random_state=1234, max_iter=30, n_jobs=-1).fit(customers_pca)\n",
    "customer_clusters = kmeans.predict(customers_pca)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare Customer Data to Demographics Data¶\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compare the proportion of data in each cluster for the customer data to the\n",
    "# proportion of data in each cluster for the general population.\n",
    "population_clusters = pd.Series(population_clusters)\n",
    "popc = population_clusters.value_counts().sort_index()\n",
    "popc = pd.Series(popc)\n",
    "\n",
    "customer_clusters = pd.Series(customer_clusters)\n",
    "custc = customer_clusters.value_counts().sort_index()\n",
    "custc = pd.Series(custc)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cluster</th>\n",
       "      <th>population</th>\n",
       "      <th>customers</th>\n",
       "      <th>customer_prop</th>\n",
       "      <th>population_prop</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>35366</td>\n",
       "      <td>198</td>\n",
       "      <td>0.001033</td>\n",
       "      <td>0.047515</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>38586</td>\n",
       "      <td>10863</td>\n",
       "      <td>0.056681</td>\n",
       "      <td>0.051842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>59705</td>\n",
       "      <td>47488</td>\n",
       "      <td>0.247782</td>\n",
       "      <td>0.080216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>56778</td>\n",
       "      <td>932</td>\n",
       "      <td>0.004863</td>\n",
       "      <td>0.076283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>4</td>\n",
       "      <td>13764</td>\n",
       "      <td>126</td>\n",
       "      <td>0.000657</td>\n",
       "      <td>0.018492</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>5</td>\n",
       "      <td>12607</td>\n",
       "      <td>12885</td>\n",
       "      <td>0.067231</td>\n",
       "      <td>0.016938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>6</td>\n",
       "      <td>25978</td>\n",
       "      <td>2981</td>\n",
       "      <td>0.015554</td>\n",
       "      <td>0.034902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>7</td>\n",
       "      <td>36688</td>\n",
       "      <td>1143</td>\n",
       "      <td>0.005964</td>\n",
       "      <td>0.049292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>8</td>\n",
       "      <td>69189</td>\n",
       "      <td>9408</td>\n",
       "      <td>0.049089</td>\n",
       "      <td>0.092958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>9</td>\n",
       "      <td>15710</td>\n",
       "      <td>14066</td>\n",
       "      <td>0.073393</td>\n",
       "      <td>0.021107</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>10</td>\n",
       "      <td>7727</td>\n",
       "      <td>1638</td>\n",
       "      <td>0.008547</td>\n",
       "      <td>0.010381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>11</td>\n",
       "      <td>39294</td>\n",
       "      <td>2424</td>\n",
       "      <td>0.012648</td>\n",
       "      <td>0.052793</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>12</td>\n",
       "      <td>673</td>\n",
       "      <td>9098</td>\n",
       "      <td>0.047471</td>\n",
       "      <td>0.000904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>13</td>\n",
       "      <td>698</td>\n",
       "      <td>963</td>\n",
       "      <td>0.005025</td>\n",
       "      <td>0.000938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>14</td>\n",
       "      <td>62671</td>\n",
       "      <td>803</td>\n",
       "      <td>0.004190</td>\n",
       "      <td>0.084201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>15</td>\n",
       "      <td>48532</td>\n",
       "      <td>15120</td>\n",
       "      <td>0.078893</td>\n",
       "      <td>0.065204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>16</td>\n",
       "      <td>24983</td>\n",
       "      <td>901</td>\n",
       "      <td>0.004701</td>\n",
       "      <td>0.033566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>17</td>\n",
       "      <td>283</td>\n",
       "      <td>11774</td>\n",
       "      <td>0.061434</td>\n",
       "      <td>0.000380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>18</td>\n",
       "      <td>28246</td>\n",
       "      <td>2407</td>\n",
       "      <td>0.012559</td>\n",
       "      <td>0.037949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>19</td>\n",
       "      <td>59991</td>\n",
       "      <td>9357</td>\n",
       "      <td>0.048823</td>\n",
       "      <td>0.080600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>20</td>\n",
       "      <td>29893</td>\n",
       "      <td>6139</td>\n",
       "      <td>0.032032</td>\n",
       "      <td>0.040162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>21</td>\n",
       "      <td>50299</td>\n",
       "      <td>6689</td>\n",
       "      <td>0.034902</td>\n",
       "      <td>0.067578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>22</td>\n",
       "      <td>24750</td>\n",
       "      <td>6525</td>\n",
       "      <td>0.034046</td>\n",
       "      <td>0.033252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>23</td>\n",
       "      <td>1894</td>\n",
       "      <td>17724</td>\n",
       "      <td>0.092480</td>\n",
       "      <td>0.002545</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    cluster  population  customers  customer_prop  population_prop\n",
       "1         0       35366        198       0.001033         0.047515\n",
       "2         1       38586      10863       0.056681         0.051842\n",
       "3         2       59705      47488       0.247782         0.080216\n",
       "4         3       56778        932       0.004863         0.076283\n",
       "5         4       13764        126       0.000657         0.018492\n",
       "6         5       12607      12885       0.067231         0.016938\n",
       "7         6       25978       2981       0.015554         0.034902\n",
       "8         7       36688       1143       0.005964         0.049292\n",
       "9         8       69189       9408       0.049089         0.092958\n",
       "10        9       15710      14066       0.073393         0.021107\n",
       "11       10        7727       1638       0.008547         0.010381\n",
       "12       11       39294       2424       0.012648         0.052793\n",
       "13       12         673       9098       0.047471         0.000904\n",
       "14       13         698        963       0.005025         0.000938\n",
       "15       14       62671        803       0.004190         0.084201\n",
       "16       15       48532      15120       0.078893         0.065204\n",
       "17       16       24983        901       0.004701         0.033566\n",
       "18       17         283      11774       0.061434         0.000380\n",
       "19       18       28246       2407       0.012559         0.037949\n",
       "20       19       59991       9357       0.048823         0.080600\n",
       "21       20       29893       6139       0.032032         0.040162\n",
       "22       21       50299       6689       0.034902         0.067578\n",
       "23       22       24750       6525       0.034046         0.033252\n",
       "24       23        1894      17724       0.092480         0.002545"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a dataset summarizing clustering information\n",
    "pop_cust = pd.concat([popc, custc], axis=1).reset_index()\n",
    "pop_cust.columns = ['cluster', 'population', 'customers']\n",
    "\n",
    "# Recalculate index\n",
    "pop_cust.index = pop_cust.index + 1\n",
    "# Sort by cluster (index)\n",
    "pop_cust.sort_index(inplace=True)\n",
    "\n",
    "# Calculate proprotions\n",
    "pop_cust['customer_prop'] = pop_cust['customers'] / pop_cust['customers'].sum()\n",
    "pop_cust['population_prop'] = pop_cust['population'] / pop_cust['population'].sum()\n",
    "pop_cust"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Compare visually\n",
    "pop_cust.plot(x='cluster', y=['population_prop', 'customer_prop'], kind='bar', figsize=(18, 8))\n",
    "plt.title('Population vs. Customer Data Comparison')\n",
    "plt.xlabel('Cluster Number')\n",
    "plt.ylabel('Proportion of persons in cluster')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 2: Supervised Learning Model\n",
    "\n",
    "Now that you've found which parts of the population are more likely to be customers of the mail-order company, it's time to build a prediction model. Each of the rows in the \"MAILOUT\" data files represents an individual that was targeted for a mailout campaign. Ideally, we should be able to use the demographic information from each individual to decide whether or not it will be worth it to include that person in the campaign.\n",
    "\n",
    "The \"MAILOUT\" data has been split into two approximately equal parts, each with almost 43 000 data rows. In this part, you can verify your model with the \"TRAIN\" partition, which includes a column, \"RESPONSE\", that states whether or not a person became a customer of the company following the campaign. In the next part, you'll need to create predictions on the \"TEST\" partition, where the \"RESPONSE\" column has been withheld."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the libraries\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from xgboost.sklearn import XGBRegressor # Extreme Gradient Boosting\n",
    "from sklearn.metrics import roc_auc_score, fbeta_score, accuracy_score, precision_score, recall_score\n",
    "import xgboost as xgb\n",
    "from sklearn.impute import SimpleImputer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3071: DtypeWarning: Columns (19,20) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    }
   ],
   "source": [
    "# Load Memory reduced train data\n",
    "mailout_train = pd.read_csv('Udacity_MAILOUT_052018_TRAIN_reduce_mem.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape before (42962, 366)\n",
      "{'D19_LETZTER_KAUF_BRANCHE', 'EINGEFUEGT_AM', 'OST_WEST_KZ', 'CAMEO_INTL_2015', 'CAMEO_DEU_2015', 'CAMEO_DEUG_2015'}\n",
      "shape after corr (42962, 276)\n",
      "Number of nan Values 0\n",
      "shape after (42962, 321)\n"
     ]
    }
   ],
   "source": [
    "# Choose the response \n",
    "X = mailout_train.drop('RESPONSE',axis=1)\n",
    "y = mailout_train['RESPONSE']\n",
    "df_mailout_train  = data_cleaning(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train,X_test,y_train,y_test = train_test_split(df_mailout_train,y,random_state=42,test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train,X_val,y_train,y_val = train_test_split(X_train,y_train,random_state=42,test_size=0.3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### XGB Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_xg = XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
    "                        colsample_bytree=0.7, eval_metric='auc', gamma=1.0,\n",
    "                        learning_rate=0.01, max_delta_step=0, max_depth=7,\n",
    "                        min_child_weight=1, missing=None, n_estimators=250, n_jobs=-1,\n",
    "                        nthread=None, objective='binary:logistic', random_state=42,\n",
    "                        reg_alpha=1e-07, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
    "                        silent=1, subsample=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[21:50:57] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:541: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "             colsample_bynode=1, colsample_bytree=0.7, eval_metric='auc',\n",
       "             gamma=1.0, gpu_id=-1, importance_type='gain',\n",
       "             interaction_constraints='', learning_rate=0.01, max_delta_step=0,\n",
       "             max_depth=7, min_child_weight=1, missing=None,\n",
       "             monotone_constraints='()', n_estimators=250, n_jobs=-1, nthread=4,\n",
       "             num_parallel_tree=1, objective='binary:logistic', random_state=42,\n",
       "             reg_alpha=1e-07, reg_lambda=1, scale_pos_weight=1, seed=42,\n",
       "             silent=1, subsample=0.5, tree_method='exact',\n",
       "             validate_parameters=1, ...)"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fit model\n",
    "model_xg.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Evaluate Model\n",
    "from sklearn.metrics import roc_auc_score\n",
    "pred_train = model_xg.predict(X_train)\n",
    "pred_test = model_xg.predict(X_test)\n",
    "pred_val = model_xg.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9084999443701693"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train\n",
    "roc_auc_score(y_train,pred_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7758407076409347"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Test\n",
    "roc_auc_score(y_test,pred_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7770589970636395"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Validation\n",
    "roc_auc_score(y_val,pred_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAGUCAYAAAA8gPUyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3debwcZZn28d9FIKKCBiQgkCjLRJFRRpDtFR0FdCSoxGVE0JFFfTMozOAyo/iCDjqD4oIgiGRQUVBHXEY0alAYVMaNkQAaxCEaWcMiQdlkEQPX+0fVCU3Tp0+d7nO60l3X9/M5n3TVU3X6rpxz6u7nqafukm0iIqJ51qk7gIiIqEcSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEwrSddIulfSH1u+tpiC7/mCqYqxwvsdK+nzg3q/biQdIulHdccRoyEJIAbhpbY3aPm6sc5gJK1b5/v3aljjjrVXEkDUQtLjJX1a0k2SbpD0b5JmlG3bSvqepN9LulXSFyTNKts+BzwJ+GbZm3iHpOdLWtn2/df0EspP8F+V9HlJdwKHdHv/CrFb0psl/UbSXZL+tYz5p5LulPRlSTPLbZ8vaaWk/1ceyzWSXtv2/3CWpFWSrpV0jKR1yrZDJP1Y0omS/gB8CVgE/J/y2G8vt3uxpMvK975e0rEt33+rMt6DJV1XxnB0S/uMMrbflsdyiaS5Zdt2ks6X9AdJyyXtP8kfc6zlkgCiLmcCq4G/AHYE/gZ4Y9km4APAFsDTgLnAsQC2Xwdcx0O9ig9VfL8FwFeBWcAXJnj/KvYBngXsDrwDOB14bRnr04EDW7Z9IrAJsCVwMHC6pKeWbacAjwe2AZ4HHAQc2rLvbsBVwKbA3wGHAT8tj31Wuc3d5X6zgBcDb5L0srZ4nwM8FdgbeI+kp5Xr31bGui/wOOD1wD2SHgucD/xH+d4HAp+Q9JeT+D+KtVwSQAzC1yXdXn59XdJmwHzgLbbvtn0LcCJwAIDtFbbPt/0n26uAj1KcHPvxU9tft/0gxYlu3Pev6IO277R9BfBL4DzbV9m+AziXIqm0end5PBcC3wb2L3scrwbeZfsu29cAJwCva9nvRtun2F5t+95Ogdj+ge3LbT9oexnwRR75//Ve2/fa/gXwC+CvyvVvBI6xvdyFX9j+PfAS4Brbnynf+1LgP4G/ncT/UazlMqYYg/Ay2/81tiBpV2A94CZJY6vXAa4v2zcFTgaeC2xYtt3WZwzXt7x+crf3r+h3La/v7bD8xJbl22zf3bJ8LUXvZhNgZrnc2rblOHF3JGk34HiKnsdM4FHAV9o2u7nl9T3ABuXrucBvO3zbJwO7jQ0zldYFPjdRPDE80gOIOlwP/AnYxPas8utxtseGFz4AGNjB9uMohj7Usn97Cdu7gceMLZSfrGe3bdO6z0TvP9U2KodUxjwJuBG4Ffgzxcm2te2GceLutAzFMM1iYK7tx1NcJ1CH7Tq5Hth2nPUXtvz/zCqHnd5U8fvGEEgCiIGzfRNwHnCCpMdJWqe8iDo2bLEh8EfgdklbAv/c9i1+RzFmPubXwPrlxdD1gGMoPgX3+v7T4b2SZkp6LsXwyldsPwB8GThO0oaSnkwxJt9tyunvgDljF5lLGwJ/sH1f2bt6zSTi+hTwr5LmqbCDpCcA3wKeIul1ktYrv3ZpuXYQIyAJIOpyEMVwxa8ohne+Cmxetr0X2Am4g2K8/Gtt+34AOKa8pvBP5bj7mylOZjdQ9AhW0l23959qN5fvcSPFBejDbF9Ztv0DRbxXAT+i+DR/Rpfv9T3gCuBmSbeW694MvE/SXcB7KJJKVR8ttz8PuBP4NPBo23dRXBg/oIz7ZuCDdEmsMXyUB8JETB9Jzwc+b3tO3bFEtEsPICKioZIAIiIaKkNAERENlR5ARERDDdWNYJtssom32mqrusOIiBgql1xyya222++NGa4EsNVWW7F06dK6w4iIGCqSru20PkNAERENlQQQEdFQSQAREQ2VBBAR0VBJABERDZUEEBHRUEkAERENlQQQEdFQSQAREQ01VHcC92Oro75ddwhT5prjX1x3CBExAtIDiIhoqCSAiIiGSgKIiGioJICIiIZKAoiIaKgkgIiIhkoCiIhoqCSAiIiGqpQAJO0jabmkFZKO6tAuSSeX7csk7VRlX0n/ULZdIelD/R9ORERUNeGdwJJmAKcCLwRWAhdLWmz7Vy2bzQfmlV+7AacBu3XbV9KewAJgB9t/krTpVB5YPNyo3Amdu6Ajpk6VHsCuwArbV9m+Hzib4sTdagFwlgsXAbMkbT7Bvm8Cjrf9JwDbt0zB8UREREVVEsCWwPUtyyvLdVW26bbvU4DnSvofSRdK2qXTm0taKGmppKWrVq2qEG5ERFRRJQGowzpX3KbbvusCGwG7A/8MfFnSI7a3fbrtnW3vPHv27ArhRkREFVWqga4E5rYszwFurLjNzC77rgS+ZtvAzyQ9CGwC5GN+RMQAVOkBXAzMk7S1pJnAAcDitm0WAweVs4F2B+6wfdME+34d2AtA0lMoksWtfR9RRERUMmEPwPZqSUcA3wVmAGfYvkLSYWX7ImAJsC+wArgHOLTbvuW3PgM4Q9IvgfuBg8veQEREDEClB8LYXkJxkm9dt6jltYHDq+5brr8f+LvJBBvRi1GZAguTnwbb5GOPieVO4IiIhmrMIyEjonnSA+ouPYCIiIZKAoiIaKgkgIiIhkoCiIhoqCSAiIiGSgKIiGioJICIiIZKAoiIaKgkgIiIhkoCiIhoqCSAiIiGSgKIiGioJICIiIZKAoiIaKgkgIiIhkoCiIhoqCSAiIiGSgKIiGioJICIiIZKAoiIaKgkgIiIhqqUACTtI2m5pBWSjurQLkknl+3LJO000b6SjpV0g6Sfl1/7Ts0hRUREFRMmAEkzgFOB+cD2wIGStm/bbD4wr/xaCJxWcd8TbT+z/FrS78FERER1VXoAuwIrbF9l+37gbGBB2zYLgLNcuAiYJWnzivtGREQNqiSALYHrW5ZXluuqbDPRvkeUQ0ZnSNqo05tLWihpqaSlq1atqhBuRERUUSUBqMM6V9ym276nAdsCzwRuAk7o9Oa2T7e9s+2dZ8+eXSHciIioYt0K26wE5rYszwFurLjNzPH2tf27sZWSPgl8q3LUERHRtyo9gIuBeZK2ljQTOABY3LbNYuCgcjbQ7sAdtm/qtm95jWDMy4Ff9nksERExCRP2AGyvlnQE8F1gBnCG7SskHVa2LwKWAPsCK4B7gEO77Vt+6w9JeibFkNA1wN9P5YFFRER3VYaAKKdoLmlbt6jltYHDq+5brn/dpCKNiIgplTuBIyIaKgkgIqKhkgAiIhoqCSAioqGSACIiGioJICKioZIAIiIaKgkgIqKhkgAiIhoqCSAioqGSACIiGioJICKioZIAIiIaKgkgIqKhkgAiIhoqCSAioqGSACIiGioJICKioZIAIiIaKgkgIqKhkgAiIhoqCSAioqGSACIiGioJICKioSolAEn7SFouaYWkozq0S9LJZfsySTtNYt9/kmRJm/R3KBERMRkTJgBJM4BTgfnA9sCBkrZv22w+MK/8WgicVmVfSXOBFwLX9X0kERExKVV6ALsCK2xfZft+4GxgQds2C4CzXLgImCVp8wr7ngi8A3C/BxIREZNTJQFsCVzfsryyXFdlm3H3lbQfcIPtX3R7c0kLJS2VtHTVqlUVwo2IiCqqJAB1WNf+iX28bTqul/QY4GjgPRO9ue3Tbe9se+fZs2dPGGxERFRTJQGsBOa2LM8Bbqy4zXjrtwW2Bn4h6Zpy/aWSnjiZ4CMiondVEsDFwDxJW0uaCRwALG7bZjFwUDkbaHfgDts3jbev7cttb2p7K9tbUSSKnWzfPFUHFhER3a070Qa2V0s6AvguMAM4w/YVkg4r2xcBS4B9gRXAPcCh3fadliOJiIhJmTABANheQnGSb123qOW1gcOr7tthm62qxBEREVMndwJHRDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VKUEIGkfScslrZB0VId2STq5bF8maaeJ9pX0r+W2P5d0nqQtpuaQIiKiigkTgKQZwKnAfGB74EBJ27dtNh+YV34tBE6rsO+Hbe9g+5nAt4D39H84ERFRVZUewK7ACttX2b4fOBtY0LbNAuAsFy4CZknavNu+tu9s2f+xgPs8loiImIQqCWBL4PqW5ZXluirbdN1X0nGSrgdeyzg9AEkLJS2VtHTVqlUVwo2IiCqqJAB1WNf+aX28bbrua/to23OBLwBHdHpz26fb3tn2zrNnz64QbkREVFElAawE5rYszwFurLhNlX0B/gN4ZYVYIiJiilRJABcD8yRtLWkmcACwuG2bxcBB5Wyg3YE7bN/UbV9J81r23w+4ss9jiYiISVh3og1sr5Z0BPBdYAZwhu0rJB1Wti8ClgD7AiuAe4BDu+1bfuvjJT0VeBC4FjhsSo8sIiK6mjABANheQnGSb123qOW1gcOr7luuz5BPRESNcidwRERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VKUEIGkfScslrZB0VId2STq5bF8maaeJ9pX0YUlXltufI2nW1BxSRERUMWECkDQDOBWYD2wPHChp+7bN5gPzyq+FwGkV9j0feLrtHYBfA+/q+2giIqKyKj2AXYEVtq+yfT9wNrCgbZsFwFkuXATMkrR5t31tn2d7dbn/RcCcKTieiIioqEoC2BK4vmV5ZbmuyjZV9gV4PXBupzeXtFDSUklLV61aVSHciIiookoCUId1rrjNhPtKOhpYDXyh05vbPt32zrZ3nj17doVwIyKiinUrbLMSmNuyPAe4seI2M7vtK+lg4CXA3rbbk0pEREyjKj2Ai4F5kraWNBM4AFjcts1i4KByNtDuwB22b+q2r6R9gHcC+9m+Z4qOJyIiKpqwB2B7taQjgO8CM4AzbF8h6bCyfRGwBNgXWAHcAxzabd/yW38ceBRwviSAi2wfNpUHFxER46syBITtJRQn+dZ1i1peGzi86r7l+r+YVKQRETGlcidwRERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQyUBREQ0VBJARERDJQFERDRUEkBEREMlAURENFQSQEREQ1VKAJL2kbRc0gpJR3Vol6STy/ZlknaaaF9Jr5J0haQHJe08NYcTERFVTZgAJM0ATgXmA9sDB0ravm2z+cC88mshcFqFfX8JvAL47/4PIyIiJqtKD2BXYIXtq2zfD5wNLGjbZgFwlgsXAbMkbd5tX9v/a3v5lB1JRERMSpUEsCVwfcvyynJdlW2q7NuVpIWSlkpaumrVqsnsGhERXVRJAOqwzhW3qbJvV7ZPt72z7Z1nz549mV0jIqKLdStssxKY27I8B7ix4jYzK+wbERE1qNIDuBiYJ2lrSTOBA4DFbdssBg4qZwPtDtxh+6aK+0ZERA0m7AHYXi3pCOC7wAzgDNtXSDqsbF8ELAH2BVYA9wCHdtsXQNLLgVOA2cC3Jf3c9oum+gAjIqKzKkNA2F5CcZJvXbeo5bWBw6vuW64/BzhnMsFGRMTUyZ3AERENlQQQEdFQSQAREQ2VBBAR0VBJABERDZUEEBHRUEkAERENlQQQEdFQSQAREQ2VBBAR0VBJABERDZUEEBHRUEkAERENlQQQEdFQSQAREQ2VBBAR0VBJABERDZUEEBHRUEkAERENlQQQEdFQSQAREQ2VBBAR0VBJABERDZUEEBHRUJUSgKR9JC2XtELSUR3aJenksn2ZpJ0m2lfSxpLOl/Sb8t+NpuaQIiKiigkTgKQZwKnAfGB74EBJ27dtNh+YV34tBE6rsO9RwAW25wEXlMsRETEgVXoAuwIrbF9l+37gbGBB2zYLgLNcuAiYJWnzCfZdAJxZvj4TeFmfxxIREZOwboVttgSub1leCexWYZstJ9h3M9s3Adi+SdKmnd5c0kKKXgXAHyUtrxBznTYBbp3ON9AHp/O79yXHPs2afPxNPnbo+/if3GlllQSgDutccZsq+3Zl+3Tg9MnsUydJS23vXHccdcixN/PYodnHP8zHXmUIaCUwt2V5DnBjxW267fu7cpiI8t9bqocdERH9qpIALgbmSdpa0kzgAGBx2zaLgYPK2UC7A3eUwzvd9l0MHFy+Phj4Rp/HEhERkzDhEJDt1ZKOAL4LzADOsH2FpMPK9kXAEmBfYAVwD3Bot33Lb3088GVJbwCuA141pUdWn6EZrpoGOfbmavLxD+2xy57UkHxERIyI3AkcEdFQSQAREQ2VBBAR0VBJABERDVXlRrCIR5AkiplbBr4K7EVR3uNKYJHtB2sMb9q1FjzsxPalg4pl0CQ9HngXRfmW2eXqWyimch9v+/a6YhuEssbZo23/sVzeHZhZNl9m+67agpukzALqg6S5wIcpSl6cC3zY9p/Ltq/bHtn6RpI+AWxK8Yt/J/Ao4JsU04F/Z/vIGsObdpIeBK4AVo2tamm27b0GH9VgSPou8D3gTNs3l+ueSHE/zwtsv7DO+KabpI8At9j+ULl8NfBLYH3gUtvvrDO+yUgC6IOk84H/BC4C3gA8C3ip7d9Lusz2jrUGOI0kXW77GZLWA24GNrd9v6R1KT4FPaPmEKeVpLcCrwTuoChyeM7YJ8JRJ2m57adOtm1USLoM2MX26rFl2zuWveIf2n5OvRFWl2sA/Zlte5Htn9v+B+ATwH9L2pZJ1jwaQqsByh7PxWW1V8o/igfqDGwQbJ9Y/qEfQVHu5AJJX5b0zJpDG4RrJb1D0mZjKyRtJumdPLz446haZ+zkX3onFN0+YIN6QupNEkB/1pO0/tiC7c8DR1Lc+bx5bVENxs2SNgCwvc/YynIo4P7aohow21dTjH2fR1H+/Cn1RjQQrwaeAFwo6Q+S/gD8ANgY2L/OwAZkpqQNxxZsnwdrro2sP+5ea6EMAfWhHAa41PaFbet3BD406mOhnUh6LPBY2yNd3E/SNhS1rRZQfOo9G/iW7ftqDSymnaS3AS8ADrN9XbnuyRQPwrrA9gl1xjcZSQB9kHSE7Y/XHUcdRmkmRC/Ki8DLKD7930nbkJ/tj9YRV90kHWr7M3XHMd3KWmj/D3hsueqPFDOgTqsvqslLAuiDpEttd50OOKpGaSZELyQdS5frPLbfO7ho1h6SrrP9pLrjGJRyGFTD+oEn9wFEr/YGdmlZvt32S8dmQtQU08DYPrbuGOoiadl4TcBm47SNDEkHdVi35rXtswYaUB+SAPqzg6Q7O6wXxaSAxw06oAEadybE2MXhUSbp5LZVpngs4Pdt/6iGkAZpM+BFwG1t6wX8ZPDhDNwuHdYJeCnFPUFJAA1x+SjP9Z/ATEkbjnV9h3kmRI8u6bBuY+DDkr5k+6RBBzRA3wI2sP3z9gZJPxh8OINVTvkG1twR/1qKD0AXAcfVFVcvcg2gD6N+s1c3ozQTYipJejTwk6b+XjRFecPjIcDbgf8BPmB7ea1B9SA9gP58pe4A6mL7o5LuAX5UTv00cDdDOBNiKtm+t3U8OEaPpMMp7ve5ANjH9rU1h9Sz3AjWn1WS5kHRFZT0GUl3Slo2UbGwYSfpFeVd0E8CngxsbfvJTT75S1pX0qHAyrpjmU6SdpB0kaTrJZ0uaaOWtp/VGduAnAI8DngO8M3y732ZpMu7XCBfK6UH0J8jgc+Wrw8EdgC2BnYEPgY8t56wBuIY4GsATamB00rSXTxyGui9wIXA3w8+ooH6BHAsxZj3Gyl6gfvZ/i2wXp2BDcjWdQcwVZIA+rN6rPon8BLgLNu/B/5L0odqjCum39OHuevfpw1sf6d8/RFJlwDfkfQ6Rr8GFqP0c08C6M+DkjanmA63Nw+fAfDoekIamO3G6e6OTYHdYdABDdg5wEgP83UhSY+3fQeA7e9LeiVFZdyN6w1t+nXo/a2ZAgy8s/wQOBSSAPrzHmApMANYbPsKAEnPA66qM7ABuJpi3nNTNflK7weBp1EMAQFge5mkvYF31xbVgNjesH1deR3kEGARxYOShkKmgfapnA62oe3bWtY9FsD23bUFNs2aPAUWQNItFAXgOrL9jwMMZ60k6ZTWOfNNMGzlYdID6FN5N2z7HZHPBt4BjHI10B/XHUDN7qXzzWDxkD3qDmCQyocjDdU5daiCXdtI2ouiy7cF8HXg/RS3gYshuyOwBxd3qokyZpjqofTo97bPrDuIGDxJr+iweiOK5yR8dcDh9CUJoD8nAAuBnwLzKcZE3237Y7VGNRg7d1g3lPVQetTxoTeS9gBeY/vwAccTg9N+7cvA74GP2f52DfH0LNcA+tA+3ifpt7a3rTOmOrTVQ/kVcJztobohph/lYyBfQ/E0rKuBr9k+pd6o6tf060TDID2A/sxq6w6qddn212qIaWA61EP522Gsh9ILSU+heCLYgRSf/r5E8YFqz1oDW7uMbE9Y0p4Uz4Perlz1v8DHbf+gtqB6kB5AHyR9lvFvfLHt1w8wnIFqq4dy/CjdHFNF+USwHwJvsL2iXHeV7W3qjWz6Sfom3R+Gs98Awxk4SS8GPg68D7iUYuhzJ4q744+wvaTG8CYlCSB6Up4AbwFW8fCTQSNuBJP0cooewLOB71BMCf2U7ZEpEzCe8j4XKH7Wn6QoB7FG+zOyR01Z8vpI279oW78DcIrt53XccS2UBNAHSSfZfkv5+sjWi7+SPmv7kNqCm2Zl6edxNaVHUN7z8TKKoaC9gDOBc8aejzDqmjjOL+lK29tNtm1tlGqg/fnrltcHt7WN9Cdg29d2+gLmUNwD0Qi277b9BdsvoTj2nwNH1RzWIDXxE2S3GzyH6ubPXATuj8Z53SidZsHUG9FglBfB5/PwC4Gftv3v9UU1/SS11vuZUZZBWPP7b/sPg49qoLaVtLjDegFDdQ0oCaA/65S//Ou0vB77Q5hRX1jTr+mzYCRtQVH86ybgMoqf+0uAEyQ93/ZNdcY3zS6h+OQ/9rt+aUubGbKTYA8WdGn7yMCimAK5BtAHSdcAD9L5079HeUZIk2fBwJoZYD9vf/avpH8EnmW7fUgwGqB8HvSr646jqiSA6EmTZ8HAhBcCl9t+6qBjGpSJnnZn+9Ju7aNM0nXlU/KGQoaA+tDkPwTb5wDntMyCeSuwmaTTaMYsmHu7tN0zsCjqsRS4gmIKMDy8B2yK2VAxBJIA+nNCy+tn8fDqkI34QyhLXn8B+EJ5cfBVFLNgzoOiTnprqewR8vhxioKJ4nmxo+ztwCspkuDZFAm/MY8F7fLBTwzZIzEzBDRFmjgfuophq49elaTPdGu3feigYqmLpK0pJgEsAK4F3m/75/VGNf0kfb9b+zBNhEgPYOokk3Y2ktNjq57gJR08qmWjbV8t6RsUjz99HfAUivsgRtowneAnkgQQ063pifFIiruDR4akbSgmACwArqcYBjrO9n21BjYg4wz9rTFMRSCTAPog6RQeOsHNkXRya3seCxiMZg9oBbAM+AZwJ/Ak4M1FVXCw/dH6QhuIbs/CNkN0I2QSQH+WtrzO4wE7G8UT4GSMYg/ofTx0XBu0tY3i8T7MKF3fyUXgKVbeDXy7G/gfK2nj9jIAndY1yShODpA0x/bKcdpeavubg45p0MqKqLfZXiZpf4q6YL8FPmH7T/VGV12KwfVB0nskbVe+fpSk71H8EvxO0gvqjW56STqm5fX2kn4NXCLpGkm7jbU1+eRf+nHdAUyDCyRt1b5S0qHASY/YesRIOhX4N+DTkj5PUQfrl8COwBl1xjZptvPV4xfFzTBjvaiFFLVhZgBPA35Wd3zTfOyXtrz+NjC/fL0r8JO64xvA8Z/U8vrItrbP1h3fNB/7vsBvgHkt694FXA7MqTu+ARz/r8p/16eogzWjXBZwed3xTeYrPYD+3O/yJw+8CDjb9gO2/5dmXV/Zwva5ALZ/RjEtcNQ1uRT4EuAw4FxJT5d0EkUhvL/2OENDI+Y+ABeznq61/UC5bODPdQY2WU06SU2HP0l6OvA7YE/gn1raHlNPSAOzTVkSVxQzoB5je6wEwlDdDdmjRpcCt32BpEOAHwA/AfZ2Q6aBAptKehvFz33sNeXy7PrCmrwkgP4cCXyV4od+ou2rASTtS1EieJS1l8RdB0DSZsBpgw9n4JpcCvwuHioH/Shgb+AWFfNAbXvUS2F8Etiww2uATw0+nN5lFlBED5pcCryqEa4DVYmkd9n+QN1xdJME0CdJM4CNbN9aLs8EDgHeavtpdcY2ncqnYb2BohLolhSfCG+kuDno07aHaiw0pt6o1oGqahiOP0NAfZB0APDvwN2SfgMcC3wOuBh4bY2hDcLngNuB9wJjF/7mUFwQ/TwwNA/F6IWkrjXfbV83qFjWYo27NtJmrT/+JID+HEPx9KcVZYnYnwIHuKiVP+p28iMferISuKi8J2DUfZuHPxaRcnk2sCkjfh2goqYPL6z1x59poP253+XjEF08/OXqhpz8AW6T9CpJa36HJK0j6dXAyI/72n6G7R3Kf59BUR/mx8AfgbfUG12sJdIDGHGtU8AANmhd9mgXxToA+CDwCUljJ/xZFDfDHVBbVAMmaR5wNLAbxQOC/jHXP9ZY60+A0+wrdQcwkVwE7oOkf+nSbNvvG1gwNZL0BIrfpbEL4euN+kmwvP/jaOAvgQ8BXxy7IaiJmlYHStKeFE9Fmwusprgz+lNjIwLDIgmgDymK9ZByDvieFHVRXmp7s5pDmlaSHqCohf9t4BEnfo9wKXBJe1DMd38QeD1FXZxtKW4A3N/2T2sMb1XNbaAAAAvBSURBVNpJOh7YDLiAYhbc1cCvgTdTPBVtrf/kPyYJoA+SlgMvsn1N2/pDgWNsb1tLYANUFn57DfByYGPgcGDxqM//Lu+CHfePxyP6FDAAST+jmAK8AfBN4GW2f1ROhDjF9h61BjjNJF1eXvcZmw59oe09ypsBf2j76fVGWF2uAfTnrcD5kva1/Rsobv6gOCE+r9bIppmk44D9geuAL1LUiF86yie+VrY/W3cMNVrP9uUAklbZ/hEUEyEkNaEO1IMtw1tbUM74sn2bxp6KMySSAPpge4mkP1EUxXoZ8EZgF4qiWCP9CZii+ulyirIP37J9n6TGdCclfZPuPYD9BhjOoLXOHnxXW9vMQQZSk/cDl5UjANsBbwKQNBv4RZ2BTVaGgKaApOcAX6coirV/E4pilXdA/w1wILAXxeyfFwBzba+uM7ZBKB8IMi7bFw4qlkGTtB/wXy3F/8bWbwu80vaH6olscCRtDGwDrLB9e93x9CoJoA8dimL9meKCYFOKYgEgaX2KcsAHAs8BLrD9mnqjqo+kL9ke6TuhozNJ29m+su44qkoCGIAmFcWStCHwiqZcC+hE0nW2u5aKGFWSTre9sO446jJsP/tcAxiMC4C1uihUL7o9F7XeyGI6lcMfHZsonhY20iSdPF4Txc2QQyMJYDCGamZAFeVzUXcAHlXW/tkA+A7wbIrnoo50MbxyymPHJkb/gTirgGt5ZB0kUdRBGnWHAm8HOj38/cABx9KXJIDBGMVxtj1tb1+O/98AbGr7AUn/DiyrObZBOKFL29CMAffoKoongD2i4qmk62uIZ9AuBn5p+yftDZKOHXw4vUsCiF6teS6qpIc9F1XSSJeBALC9Z90x1OgkYCOKe0DajfwMIOBvKX//29neesCx9CUJYDBGbgiIEXouai8k/XW3dtv/PahYBs32qV2aFw0skJq01jcq5/5je1V9EfUus4CmWFOKYk1QCA/b7x1ULHUobwRrZ+CvgDm2G/M8gAbWgRLwL8ARFB941qEoCHfKsBWATA+gD5KOsf1v5evtKW4GW6/8BXm17f+Bh39iGCGf7lYIb9DBDJrthx1jeTPg0cBNFCeGkTdOHah/rjWowXgLsAewi+2rASRtA5wm6a22T6w1uklID6APrc/8lPRt4OO2z5W0K3CS7WfXG+H0SSG8gqS9gXdTfPp/v+3zaw5p2nWoA3UORR2ooRr/7pWky4AXjpU/b1k/GzjP9o71RDZ56QFMnS1snwtg+2cNKIrV2EJ4AJJeTPGJ/w7gaNs/rjmkQWp0HSiKYni3tq+0vUrSUE0BTgLozzaSFlOMA86R9JiW+ihD9YswWQ0vhAdFGeSVwO+Bd7YXgRzxYnBP5KE6UCdJ+j7waEnrNqEOFHB/j21rnSSA/ixoW14HQNJmFJ+ORprtC8q6+D+gKIS3dxMK4ZUaOw20nPJ7LkXyH6sD9RjgBklNqAP1V5Lu7LBewPqDDqYfuQYQPUkhvEJ5AvwLiv+L3zYoAT5C6kANnySAPpRPA3oDxWPhtqQ4CdwIfINilszI3xDVVOXP/v0Uj0S8lqL3Nwf4DMU1gZH+2XerA2W7U4mEWAslAfRB0heB24EzKcaDoTgJHAxsnJLAo0vSicCGwFtt31WuexzwEeBe20fWGd90aq0DRfEs3NY6UDNsj3QdqFGSBNAHScttP3Wctl/bfsqgY4rBkPQb4Clu+wMqH5Rzpe159UQ2/ST9apw6UAKWjT0vN9Z+60y8SXRxm6RXSVrz/yhpHUmvBpowE6bJ3H7yL1c+wGgW/2u1pg4U8LA6UBTXgmJIZBZQfw4APgh8QtLYCX8WxeMRD6gtqhiEX0k6yPZZrSsl/R2jXw200XWgRkmGgKaIpCdQ/H/eWi6vN+oXAptM0pbA14B7gUsoPvXvAjwaeLntG2oMb1o1vQ7UKEkCmEJNK4oVIGkv4C8pPv1eYfuCtvaRexyopDnd6kDZ7lQoL9ZCSQBTYJyiWItH7Q8/Jq+1XtSoSB2o0ZGLwH2QdFw5G+T9wOXAjsAq22fm5B+lUXwWxFgdqDUznco6UG+jAXWgRkkuAven6UWxYmIj9/uQOlCjIz2A/jwROA7YD1gh6XOURbHqDStiepXXOg6hqAO1DUUdqJz8h0yuAUyRlqJYBwLPAZpQFCsmIOmyYaoPX0XqQI2OJIBpkKJYMWYUHwcaoyMJoE8pitVMLZ+C4aELvaa4rjbTdoYBY62XX9I+tBbFktReFOsMIEWxRpTtDVuXy17fm4G/p3hEYsRaLz2APqQoVkiaRfGQ8IOA/wBOtP37eqOKqCY9gP6sKYol6WFFsSSlDMQIk7QJ8Hbg1RS9vR1t31FvVBGTkwTQnxTFaq5rgVUUD4C5B3hD63OBbX+0prgiKksC6M8nKR4K0v4a4FODDycG6MM8dBF4w24bRqytcg2gDymKFRHDLHcC9+cCSVu1ryyLYp008GhiYCR9ueX1B9vazht8RBGTlwTQnxTFaq7WRz6+sK0t139iKOQaQB9SFKvRuo2dZlw1hkISQJ9sXyDpEIqiWD+hKIp1X61BxSA8RtKOFL3oR5evVX49utbIIirKReA+pChWc0n6AV0+6dvec3DRRPQmCSCiB92e+Sxpa9tXDzqmiMnKReCI3iyWNLN9paQdgO/XEE/EpCUBRPTmEoqL/48ZWyHp+cAS4P/WFVTEZGQIKKJHko4G9gHmAy8CTqR4DsTSWgOLqCizgCJ6ZPs4SfdS9AYE7GV7Rc1hRVSWHkBEDyR9k4dmgO0BrABuHmu3vV9NoUVUlgQQ0YPySXDjsn3hoGKJ6FUSQEREQ+UaQEQPJH2f8W8Es+29BxlPRC/SA4jogaRndVi9O/AO4Bbbuww4pIhJSwKI6FN5PeDdFOVA3m/73JpDiqgkQ0ARPZL0IooT/33AcbZzB3AMlfQAInog6WKKuv8fBn7a3m770oEHFTFJSQARPWirBjp2P8AY295r4EFFTFISQMQU61YpNGJtkmJwEVNAhb0kfQpYWXc8EVUkAUT0QdJukj4GXAssBn4IbFdvVBHVZAgoogeSjgP2B64DvgicAyy1vXWtgUVMQqaBRvRmIbAcOA34lu37JOXTVAyVDAFF9OaJwHHAfsAKSZ+jeDh8PlTF0MgQUESfJK0PvAQ4EHgOcIHt19QbVcTEkgAippCkDSmeCnZm3bFETCTd1YgelTWAbrO9TNL+wF8DvwU+UW9kEdWkBxDRA0mnAjtQFID7NbAB8B3g2cAM26+tMbyISpIAInog6Ve2ty/H/28ANrX9gCQBy2w/o+YQIyaUWUARvbkPwPZ9wLW2HyiXDaQMRAyFXAOI6M2mkt5GUQRu7DXl8uz6woqoLkNAET2Q9C/d2m2/d1CxRPQqCSBiGkl6l+0P1B1HRCe5BhAxvV5VdwAR40kCiJhemniTiHokAURMr4yxxlorCSBieqUHEGutJICI6fWVugOIGE9mAUX0SNKewCuBucBq4DfAp2yvqDWwiIrSA4jogaTjgYOAiyju/L2KohDcVyRl5k8MhfQAInog6fKxej/lQ2AutL2HpI2AH9p+er0RRkwsPYCI3jwoaePy9RbADADbt5ELvzEkUgsoojfvBy6TtBzYDngTgKTZwC/qDCyiqgwBRfSo7AFsA6ywfXvd8URMVhJAxBSTtJ3tK+uOI2IiSQARU0zSdbafVHccERPJNYCIHkg6ebwmYNYgY4noVXoAET2QdBfwduBPHZpPsL3JgEOKmLT0ACJ6czHwS9s/aW+QdOzgw4mYvPQAInpQzgC6z/Y9dccS0askgIg+lXP/sb2q7lgiJiN3Akf0QIVjJd0KXAn8WtIqSe+pO7aIqpIAInrzFmAPYBfbT7C9EbAbsIekt9YbWkQ1GQKK6IGky4AX2r61bf1s4DzbO9YTWUR16QFE9Ga99pM/rLkOsF4N8URMWhJARG/u77EtYq2RIaCIHkh6ALi7UxOwvu30AmKtlwQQEdFQGQKKiGioJICIiIZKAoiIaKgkgIiIhvr/D3Wj3sJgJwkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate feature importances\n",
    "importances = model_xg.feature_importances_\n",
    "\n",
    "# Sort feature importances in descending order\n",
    "indices = np.argsort(importances)[::-1][:5]\n",
    "\n",
    "# Rearrange feature names so they match the sorted feature importances\n",
    "names = [X.columns[i] for i in indices][:5]\n",
    "\n",
    "\n",
    "# Create plot\n",
    "plt.figure()\n",
    "\n",
    "# Create plot title\n",
    "plt.title(\"Feature Importance\")\n",
    "\n",
    "# Add bars\n",
    "plt.bar(range(5), importances[indices])\n",
    "\n",
    "# Add feature names as x-axis labels\n",
    "plt.xticks(range(5) ,names, rotation=90)\n",
    "\n",
    "# Show plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Part 3: Kaggle Competition\n",
    "\n",
    "Now that you've created a model to predict which individuals are most likely to respond to a mailout campaign, it's time to test that model in competition through Kaggle. If you click on the link [here](http://www.kaggle.com/t/21e6d45d4c574c7fa2d868f0e8c83140), you'll be taken to the competition page where, if you have a Kaggle account, you can enter. If you're one of the top performers, you may have the chance to be contacted by a hiring manager from Arvato or Bertelsmann for an interview!\n",
    "\n",
    "Your entry to the competition should be a CSV file with two columns. The first column should be a copy of \"LNR\", which acts as an ID number for each individual in the \"TEST\" partition. The second column, \"RESPONSE\", should be some measure of how likely each individual became a customer – this might not be a straightforward probability. As you should have found in Part 2, there is a large output class imbalance, where most individuals did not respond to the mailout. Thus, predicting individual classes and using accuracy does not seem to be an appropriate performance evaluation method. Instead, the competition will be using AUC to evaluate performance. The exact values of the \"RESPONSE\" column do not matter as much: only that the higher values try to capture as many of the actual customers as possible, early in the ROC curve sweep."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3071: DtypeWarning: Columns (19,20) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    }
   ],
   "source": [
    "# Load Test data\n",
    "mailout_test = pd.read_csv('Udacity_MAILOUT_052018_TEST_reduce_mem.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VHN</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1754</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1770</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>1465</td>\n",
       "      <td>2</td>\n",
       "      <td>9.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>1470</td>\n",
       "      <td>-1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1478</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 367 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0   LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  ALTER_KIND2  \\\n",
       "0           0  1754         2         1.0       7.0          NaN          NaN   \n",
       "1           1  1770        -1         1.0       0.0          NaN          NaN   \n",
       "2           2  1465         2         9.0      16.0          NaN          NaN   \n",
       "3           3  1470        -1         7.0       0.0          NaN          NaN   \n",
       "4           4  1478         1         1.0      21.0          NaN          NaN   \n",
       "\n",
       "   ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ...  VHN  VK_DHT4A  \\\n",
       "0          NaN          NaN                   6.0  ...  4.0       5.0   \n",
       "1          NaN          NaN                   0.0  ...  1.0       5.0   \n",
       "2          NaN          NaN                  11.0  ...  3.0       9.0   \n",
       "3          NaN          NaN                   0.0  ...  2.0       6.0   \n",
       "4          NaN          NaN                  13.0  ...  1.0       2.0   \n",
       "\n",
       "   VK_DISTANZ  VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008  WOHNLAGE  ZABEOTYP  \\\n",
       "0         6.0      3.0             6.0             9.0       3.0         3   \n",
       "1         2.0      1.0             6.0             9.0       5.0         3   \n",
       "2         6.0      3.0             2.0             9.0       4.0         3   \n",
       "3         6.0      3.0             NaN             9.0       2.0         3   \n",
       "4         4.0      3.0             3.0             9.0       7.0         4   \n",
       "\n",
       "  ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0         1                    4  \n",
       "1         1                    4  \n",
       "2         2                    4  \n",
       "3         2                    4  \n",
       "4         2                    4  \n",
       "\n",
       "[5 rows x 367 columns]"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mailout_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "mailout_test = mailout_test.drop(['Unnamed: 0'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>...</th>\n",
       "      <th>VHN</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1754</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1770</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1465</td>\n",
       "      <td>2</td>\n",
       "      <td>9.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1470</td>\n",
       "      <td>-1</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1478</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 366 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  ALTER_KIND2  \\\n",
       "0  1754         2         1.0       7.0          NaN          NaN   \n",
       "1  1770        -1         1.0       0.0          NaN          NaN   \n",
       "2  1465         2         9.0      16.0          NaN          NaN   \n",
       "3  1470        -1         7.0       0.0          NaN          NaN   \n",
       "4  1478         1         1.0      21.0          NaN          NaN   \n",
       "\n",
       "   ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ANZ_HAUSHALTE_AKTIV  ...  \\\n",
       "0          NaN          NaN                   6.0                  2.0  ...   \n",
       "1          NaN          NaN                   0.0                 20.0  ...   \n",
       "2          NaN          NaN                  11.0                  2.0  ...   \n",
       "3          NaN          NaN                   0.0                  1.0  ...   \n",
       "4          NaN          NaN                  13.0                  1.0  ...   \n",
       "\n",
       "   VHN  VK_DHT4A  VK_DISTANZ  VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008  \\\n",
       "0  4.0       5.0         6.0      3.0             6.0             9.0   \n",
       "1  1.0       5.0         2.0      1.0             6.0             9.0   \n",
       "2  3.0       9.0         6.0      3.0             2.0             9.0   \n",
       "3  2.0       6.0         6.0      3.0             NaN             9.0   \n",
       "4  1.0       2.0         4.0      3.0             3.0             9.0   \n",
       "\n",
       "   WOHNLAGE ZABEOTYP ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0       3.0        3         1                    4  \n",
       "1       5.0        3         1                    4  \n",
       "2       4.0        3         2                    4  \n",
       "3       2.0        3         2                    4  \n",
       "4       7.0        4         2                    4  \n",
       "\n",
       "[5 rows x 366 columns]"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mailout_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape before (42833, 366)\n",
      "{'OST_WEST_KZ', 'D19_LETZTER_KAUF_BRANCHE', 'CAMEO_INTL_2015', 'CAMEO_DEU_2015', 'CAMEO_DEUG_2015', 'EINGEFUEGT_AM'}\n",
      "shape after corr (42833, 276)\n",
      "Number of nan Values 0\n",
      "shape after (42833, 321)\n"
     ]
    }
   ],
   "source": [
    "# Apply the preprocessing step\n",
    "mailout_test_clean = data_cleaning(mailout_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make prediction\n",
    "testpred = model_xg.predict(mailout_test_clean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "# prediction_for_kaggle = reg_model.predict(mailout_test_clean)\n",
    "df_kaggle = pd.DataFrame(index=mailout_test['LNR'].astype('int32'), data=testpred)\n",
    "df_kaggle.rename(columns={0: \"RESPONSE\"}, inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Kaggle submission csv file\n",
    "df_kaggle.to_csv('submission1.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For first try I got a score of 0.79473 but I will try to make some improvement"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAA68AAABvCAYAAAADmPbAAAAgAElEQVR4Ae2d608cV7qv+y/Ix/jL1mg80pb2Ho1GE0WjaCvSKPKJRtl8OXHyJUbRliMfTRxZcpJxdOwzTjy54InjMJkE47EnECfBIXaISWLPtB3bQCAeMIQAxiENAaahjdsNaS5taHMRQr+jd1VV96rq6hsYKOBnqdW4uy5rveup6nrqXWuVLzw1C74YAzJABsgAGSADZIAMkAEyQAbIABnwEgO+wzWQl/y755574PNS4VgWHixkgAyQATJABsgAGSADZIAMkAEyIAxQXplpZqadDJABMkAGyAAZIAN5MXBrahYjUzMYnZrGxNQdxKbiuD0Vx392vISt3X/Frr6PUTR4Hud/7MYPt2N5bZuSQkkhA2QgHQOUV/5Y8QeFDJABMkAGyAAZIAM5MRAxhVVE1e31b23/F87Xf3YcwDs36iDrprsg5eeUFTJABnJhgPLKHyv+kJABMkAGyAAZIANkICsDkmXVhXV86g5+nJrG8NQMJBMrF549t2P4+0g3ysJNeG3wPP5X158TMvvA9T/h/XBT1v3kcgHLZSg6ZGBjMkB55Y8Vf0TIABkgA2SADJABMpCWAcmYStdgS1xFYvPJov5jpBv/O/DXhMQ+3PUWem9PpN0fpWRjSgnbne2eCwOUV/5Y8ceDDJABMkAGyAAZIAOuDEhW1RrPKgIr/09cYN6+g/DYBMIjPyJ8K4Jbt27hlrwPD6vPbkbHcTN2J7G8f6Q7kYkVge2jwCZik4gpOWRMyEBGBiivBCQjIDyZ8i4YGSADZIAMkIGNyYBkVy1xHZtKSmj49h3cio6asirCmuUVHYWsIxxJt2KrKzEzsBuTK55P2O5LYYDySnmlvJIBMkAGyAAZIANkIIUBq6uwTVzHb2eX1TQyGx6/rfbhFNilXMhyXYoQGdhYDFBe+WOV8mPFk8DGOgmwvdneZIAMkAEy4GTAmpxJBDbx3diEq7gODQ0hGAyiv78ffX196jU4OIhwOJyyvOpm7MjAnuAkTskY87qUsSADGRmgvBKQjIAkfrAYJ8aJDJABMkAGyMCGYEC6C1uTMyXGuKbJuIq4/vDDD2htbVWvr7/+GjU1NWhpaVEy6yqwZgZWxsDKY3VkFuJ8JoDitQlvtpCBjcsA5ZU/xBvih5gnuY17kmPbs+3JABkgA/kxYGVd5V3FTsa4unQFFnHt7OxEb28vxsfH1TIisW1tbbhw4YJ6DQwMuK5rjYG1ZiFm9jW/NiLTjNdGZYDySnmlvJIBMkAGyAAZIANkQDEgz2u1sq5WNtRtcibJqH7zzTc4d+4crly5gm+//RanTp3Ce++9h4aGBtTV1eHMmTNoampylVfZplx8y2N0rOyr9azYjXpRznpTSMlAdgZWVV4Hh8cwGEtXyCn09XSioWc4wzLGupm3k277Hvg8Oozum26v2N29iJiIqf0EzR/m7kvHcbCqB8b/ze8m3OMRTClfatnUMlH39VfsIFSxTC3biu2fFz13l1nGk/EkA2SADKwKAyNml+Fxa6yrI+t68+ZN/Otf/1JSKlnVa9euKUm9fPky/vnPf+Krr75S3YVFbE+ePIlPP/3UdeyrZHKt7OvBQT8+i3wL2ff6+t2eQl+wO6dr2fD4GPqG3V+DzmMhNoVATyeagmNI+c657NSU2q7tejtmfOa2P9ty5rYGh0Noau7GtdW+1kup2ypfe+ZbnugY+tZiDG/fwcCYNvZdq/eNH4fRGYpiIK3P3f02Wh15He7HmXeex8OPFOBYu0ulgvV47alC/M8Le7H7hd/h8adeRcV3U6kntGzb0YLrxZNhx4lH4fP5Ul9bP0DH3Sx73cvw+R5FaY/EOoSybT74fv026mUfPR+gwOfD3jqXdpgKoXRravnu/c1zOPm9tXwD9vp8KDgRSm2fu1mHLNsyYvkyzmVZzoscsEwWS3wnC2SADJCB1WbA6jL8o9VlWJukSSZhkm7CIqwiqoFAQI1zlQyrZFtDoRAikYjqQixdiSUre+LECbWOSK+z67E1eZPsS7K9iW7K6+G3/FYrSl4oxLbd1rXsWzgbTM934MKr6rpYro3tr1dxZjC53uB3n2DPE+Y18u7tePipzNu99vmLqdfb7R859pHcp+26PNqNile24/Hf7cXuvXux86lC7K7ozEGYk+VdbZ69sf8xtF44gp3Stu93rur1cr7xGOhvxakPjuKNkmZ06sfleBg1n59AycnP8OGZT3G07DNcDLoLbr77zLb8ystr3z+w56m9+POFRlT80U1eQzj1/wrx0oWkDPU1n8CO3Z+jVQ9a1u14/8AxhGsfTuaQ3czWkBm/t8nrLMKSibXu/OQir3u+SGaIv/8Cu3/tg++3x3FVtQflNWPsdWb595o6YbNdvX8OZRuxjcjA3WfAejyONVHTrZGoks4bN24oCZXxrCKwly5dUgJbW1uLyspKnD9/Xk3cJII6MjKC7u5uNXGTdCX+4osvIOs75VW2LW0o+xJ5tc1svKZ/M8dw/i+F2FPdn/jdu3bhdTy++xM05VGvvq+O4PFXLiNgrTPeipInnsex1rHEdvu+Po4d6bYbvIx9z7yKfXvdrrcd7Hz3CXY+8xEaxpOfN7xfiB3lreiz9h/tRMkzu/Ded8lleAxmi0UIp17cjt2ll3G+6tU1JK+TaLvyGY5++jWa2xsd8noHnVdOo6whiAGTjRvBZpwo+xptGj/LxcbKy2ugFeeHpKGHceYVl4NpvB9nK/+BhmEdhk6UPPIWzt7SPsu2HetA8/B7TtnC/gaUvvkctm3ZjqeLjqOqPxmDq1WHcOBYA662nMbeJx/C01Wm8N+8hrKinSjYuhMHPr6GoENerfW6JTa5yOuLDYmTpIBoL7clrz3o8B/H008+hG3P2svZXXccB4q+MGXXKL+tDKqNhlH/8SFs3/oothedRv1wD04WHUJx3XBy3xliYS+TsY/ultM4+Ox2PChxONGADmfXaImTiq3s8wOc02K7XAcct5vkl7FgLMgAGSAD3mMgNhVXImmNP70VGVbSKY/CaW5uVt2CZVzr+++/j/Lycrz22mt4+eWXcezYMZw9e1ZlYGU56TYs414/+OADvPvuuyormyKvEeM33hpnK/teF0wMXsa+R47jsu1CvhvvPVPo3uPQ9Vq1G+/t3oVj7VrPw9YTePiPl3HNtnwIp/a6CWUIZ14pxJ8b+92vt23bMGR73wXtmmtqCoFrrWjVr72nxnC2uAD25bzHsLcY6sflr424quz6msm8TqAtEDbkNNhsl9dYEH8vOYeGsN72k2i+cBRfBCYTx/BAuA8NdZfx4bk6nG+/kRDdpbbPystr4mBJI6+J77WAfPcJdjyT7m5VHttx2/YqfuYmXLYGbTyE+32b8OD2QyitPo2DTz4A36YnUKa6/87i3Is++H75K9z/8yfwdNEhHPSHEB7+Ert/5sO9v9mO3UWHsHv7I9i2TbonW92GzfWsrsl5y+swTj67KSXz+uBvH8WW7ftwoGgftklmdtNOnDS7ubjVU5XdKsPUME7u/Cl8mx7Ctr2HcGDvdmx58gnVnTnRHTlLLJz7uHriCdzr+ykKnj2UKNO9v3kbF632/v4DbN3kw0+2PocDEicV24dwsFHjzlqW74kTkY1PxoVxIQNkgAysOwasyZqs870unDK7sHQNli7A0oX44sWLeOWVV/DMM89g165deOGFF5TQfvbZZ6iqqlIZWRFdEVtZT9+W9be1H+d+rc/X4vtg43E8rGdMzeOkoTx38UvJuso2RF5TthvCmT+mbjdw6XVs+0sj+tIli/Rj1yXr6hr3W/U49MSLONXHayXX+Ogxdfnba/J6IzqMQCSHrr5OeQ21oqysGd856vjDNT/euHLDOCeGWnGi7BwudoXxQ+QGGi5/ihNXbuCGY53FxNH78hrtxLHdhTj0VbKLhL2ia11ef4r7tzyEB7XXwQbjpBDsaUaV/xpUhlQ1dgN2+3zYWmFkWJUAbnoOVVqW+uqRh+CzfRbDub2/Wpq8/uyBZPl+vgn3bz+Oi4l9GpnXe5/90pwAahbhzuPY4vPh6bPGBEpOsZT2s8lr4yH8wrcpsbx8H7y0D7/QxtJmi4VtH9EvsXuTD1uO9CQvKgZPY5vPh20fy92vGKr26AIu8R5G2ZM++J48rcWbJ2f7scZ4MB5kgAyQgfXOgFMiLcl0exeZlcmb5PE4Iqv79+/H73//e7z00kvYu3evej906BDeeeedDSWv6SSl4f0CPF7Rnbw2SXshL1laR9ZVlr1Vj9ceeREVZhJDWBxs/0iNpbRlQ2814tBTR3BeZU2zXSeP4ewbqfKrc95UJWNen8e2Ha+ioj3d9TjPDXrM3P5Ox4Xbssv9Wf93X6Lkg0/x4akKHL3QiYD0EhjvwxfOsa3CnVNe5f/nvke/g9/+Tj/eqAsqQVV/N5giK8vFJtEfzUGUHdt0i4PH5TWEM68XYuf7mQaHZzsovXswGcL1CLYXSXYw+TrZqZU5Ooz6ui9R9fFxHHj2UfxEEzqbAKrGNidj0kVSPr+0b2nyuv1tVPm/NF4fv60yqz958nSGMa9WV2JDsm1iaUKpl72j4gn4fM+hyhqHq5YxRD2ReZXPMsTCto+Gl3Gv7yEU63GU7O7/8cGnYtOMvZt8eFCX26lZdFftdCmH1hZm2d0OJH7GOJEBMkAGyMB6YCBdt2E3eZXP5JE5Mp5VJm/66KOPVOb1ww8/RFlZGf72t7/h8OHDeOutt1SmNmUb67TbsJKU8tSJeXKV177at1wyrMbxde3SW9j2xO/w0rFPUPKXF/E/L3yEY+/o8jmFy8e246VLVhfgLNfJqnejfayrk+O+nlZcbm7F2aq3sDPLBFHOdfn/5HnRO/I6gcZzp1ETkrLdwXctfhwVkT15Aicaw6nZ0Xzk1RJWybyWVODUle/xzV2ejdjD8hrC+XeeV+KaGCjuKg9ZDkrXdZIgreZBZRMul3LKWNUHpWvrr2W86yEUizhmkVc1O7BjjGrYMeZVF8fFjHkNt72NB32bzBmK7aJqxNP+mVs99TK4fR+esm8jWyxs23DU12rjc3t98G07jQ7Htq3vDcl/AmUc+5rDXWFvHEOJtnM5fvgd24gMkAEykD8D6SZsShHPW7ds3YBFYFtaWlQXYRkPW1paqqT1yJEj+MMf/gB5rE7KNtbrhE2uY1NFKnXJTNc2abKu+u/crX40NLficnsIfVPGdq1ZgqXL8uMvfo6mW9Zjd/px6o8FKPna7TEt2bOuzmNIzV58rJUzDuvtkePf3pFXF/bGJ9M+Dicl8xruxIcuGdofvv0MJY3h5DXkj2E12dMX5z7F0ZIKVH9nTNDmZCrf/3tUXqfQ8P7zePx156B0l2Dn0pc/R6jyDd5Sl7cJV0oZzcfU7NUnS7JnI3UBtMpy7sVN8G2xZgI24nX12CNLy7w6ZfjmF9ie6L5sl0yjHPbPjHru0x5jE8PJnT74rDGvdZIp/RUO6ONNlSBbj+DJHgtbLNU43k3Y7def+9qMA7/04RevN6vHBYnk37tH6+o8NYuLRb+C75eHkuNiU9rEjT9+ZrHHd7JABsgAGVj7DGR6VE6KfGoCKxlYmdTpr3/9KyoqKvD222/j9ddfx8GDB1UXYjd5XbePylHdex0TjU4ZUpptpt5MWddwsBHvVbUmZx+W6xQ1A/ERnDeHc0l21/6oHe3/zsmCMmZdQzgvE6jaJmyaRbrxvDz2sx/7npbXTNe8zszr1A1cLLMyt1a9o2j4zD5hk42JG+34sKQO34xZyy/+3YPyOoXWz1/HtpzEVSq+XjOvZjfXrcdxVbrTTgzj4hGZhMgSOse4URM6yVDK+NGtR5rRcXMYHY3Hse1nm5Ymr3q3Yf9pHNj6U217dlE1QHV8psa0+nD/ntOo7+/BuTe34983afI61YPi38qzZ3eiuPpLVFUfx/atD+D+RF2zx8Imr6qL8Cb4fr0PJ7+PITgcMmP3SKIrsXQRvtf3AHZX9yA4HFNxkgmcbONkMx3I/C55Z42xYCzIABkgA+uGgRHzsTXjU+b4tNt3UjOmmrTqQivZV5ld+OOPP0ZJSYl6SZfh4uJi127D4dvGPmRfMtZW9m274F2zXEk2tBA7/9aIQEyuVcdUUsY22VKwFe+V/wNNtiFTWbKu45049kwhXpIJOiU2sWGc/9vz2FnRnSETmu46OVvW1ZiBWNXBmjX5ViuO7c4le7x4MVkf7e9e//Ujr7P44dtzKPF3ImDK6A/XvkTJqWZ8p3ifReeVCtsETQO9jWqSJ9uzYhd5fHtPXtX04tpdIu1hzVaXCDvY6Q5Kd3Ds667uMnbhcinL96ex7ecinj74fD/FtneP4+mE0LnLq0xGVH/MlEOfzDr8Ms5Vv6zJpmO9XGYbVvuXMshrE/79v59DaZuV1XSIqgLR+VkMF980xutKPbYUNeCkzJRsZV5lneFrKH32UWNiqK3P4eT3jm1kiUVKLIevoXj7fyjZV+X+2aM42GCN/5BY2+Mk5Sp4s5mTNS3yROKl44plcTmXsF3XyQUx25bH9/IzYD22RmQyYsrkrehoTgIrMwrLuFeZoEm6DUv3YRn76jrmNTqqjkvZhzVJlPV4nnXRztF+nCr+HR43r2O3vfI5mhKTXc5isPkEHn/keVQEkm2aMetqnceD9fjzC4VmdrUQu0sbHY/OSW7PiGOa6+SMWVdzG9F+nCndm6jDwzLWtiqTKDv3zf/rLK8neQ1PRfFN3TkcLTmKN0qOouTTr/FNRGvv8TC+OleBN8yxtG984MdXoTU/YZNWQeuA5LvrBVbw5jC6bXfmcojdRAzdNy3BzGH51Y79RCw5W7Eqi31mZevgzzsW0eEscZA4DTv2vQbitdrtxf27HqsWp3znMUQGyMBaZsDqOizvqh45ZF/l8Tl9fX3qObBnzpxRY1+lC/Gbb76pMrDOR+VYWdeUfa2335dxt7Gmd+H4iI5h0MxyLTtrsSn0DWvPm11vbcT6LP6a5vYdDFiZebc4ZhpL67Z8Dp+tYub1Lhy4OVRw2Q9olmHxwFuxk2fT/tzqVjyM7v5rOPnyI7hXezYt25HHCxkgA2SADJCBlWFAz4YOW115x29nzL6KnDY1NeHatWs4e/YsZMbhd999V2Vg5V2X1/D4bXXtINu2sq5WlpdtvDJtzDgzzmuVAcqrJVB8X7qELiWG35/G0/9tdfM1uya36N18eZJZqycZlpvskgEyQAbWHgNWRlRmH06039hEWoEdHBzElStX0N3djfPnz+Pzzz9XAnv06FGVfZXvZXysNUmTbNOa2TiR4V3KdQTXTbYTY8FYrGMGKK/ruHETPzasI09iZIAMkAEyQAbIQJ4MWHI5pgusSwZWZhoWOe3o6FDPfG1sbMSFCxfwySefqHGvf/rTn9T3VsZVrk9km5J1tclxnuXjdc7auynCNmObLZUByitPlPwxJwNkgAyQATJABshACgPSlTc2FVeSaRNYGQOrTeI0NDSkxruKwI6OjqK/vx/19fXqkTkyWdP5y7W4GUtmcC1xlW2zuzBlZqkyw/U3FkOUV/5YpfxY8SSwsU4CbG+2NxkgA2SADKRjQMalWgIrWdLEGFi5fpJH3YxNYGDoJlq/bUM0GkV0dBQj0Sj6B0No6+rGD+GRxHWGrGtlc2Wbtm3xeiwRp3Rtwc95nJKBWVBeebLkyZIMkAEyQAbIABkgA2kZkOyoJZ3S1VfGqOaTMZVlrTG0VlfhfNbnBTuljQyQAYsByit/rNL+WFmQ8J0nDDJABsgAGSADZEAXUJHQ8ak7+HFqWmVQ9We0yt+SVZXvZBlZ1npxciZyxHMJGVgKA5RXyivllQyQATJABsgAGSADOTHgzKJaUprtPd9s7VIubrku5YgMrF8GKK/8scrpx4ongfV7EmDbsm3JABkgA2QgXwYkuzpidgeWLsXWuFiRWPlbPhNhlWX0rGy+++HyZJMMkAGdAcor5ZXySgbIABkgA2SADJABMkAGyAAZ8DwDlFdC6nlI9bst/Jt338gAGSADZIAMkAEyQAbIwMZkgPJKeaW8kgEyQAbIABkgA2SADJABMkAGPM9AirxaH/C9JvEcIcaCsSADZIAMkAEyQAbIABkgA2SADHiHgXvuuQc+Noh3GoRtwbYgA2SADJABMkAGyAAZIANkgAykMkB5PZwaFILCmJABMkAGyAAZIANkgAyQATJABrzFAOWV8squ0mSADJABMkAGyAAZIANkgAyQAc8zQHklpJ6HlHe8vHXHi+3B9iADZIAMkAEyQAbIABlYDQYor5RXyisZIANkgAyQATJABsgAGSADZMDzDFBeCannIV2NuzrcJ+8mkgEyQAbIABkgA2SADJABbzFAeaW8Ul7JABkgA2SADJABMkAGyAAZIAOeZ4DySkg9DynveHnrjhfbg+1BBsgAGSADZIAMkAEysBoMUF4pr5RXMkAGyAAZIANkgAyQATJABsiA5xmgvBJSz0O6Gnd1uE/eTSQDZIAMkAEyQAbIABkgA95igPJKeaW8kgEyQAbIABkgA2SADJABMkAGPM8A5ZWQeh5S3vHy1h0vtgfbgwyQATJABsgAGSADZGA1GKC8Ul4pr2SADJABMkAGyAAZIANkgAyQAc8zQHklpJ6HdDXu6nCfvJtIBsgAGSADZIAMkAEyQAa8xQDllfJKeSUDZIAMkAEyQAbIABkgA2SADHieAcorIfU8pLzj5a07XmwPtgcZIANkgAyQATJABsjAajBAeaW8Ul7JABkgA2SADJABMkAGyAAZIAOeZ4DySkg9D+lq3NXhPnk3kQyQATJABsgAGSADZIAMeIsByivllfJKBsgAGSADZIAMkAEyQAbIABnwPAPLKq/3VXeh0H8dW8pSjX3z6eso9HehoCL1u+W4w2HtT/Zpe51tw+bDNShoGUX7UBg7lhXaFhQ492/+f6XikFdsq/tR2RPB/uqVaaO8yras7cT6si3IABkgA2SADJABMkAGyIDXGFhWed3cOokZABMDvXaLLxtA0xyA2XHsWSEJ2ROah+s/swzlUfl2DtX+xUO642oElR292JK2TmEEXAsBBK7nut8AiroiKK1vs8c07T5z3a5jueohtM+ahZ2dRCkFdnnjfbfbj9tje5EBMkAGyAAZIANkgAysMwaWVV59h9tQPiYCNI1qTX529Iu5LqCpqXHFgDLkdRrVzsynmXn1lbXhseqWJZVH1TWjkBvyOhEO2bO/eWWgzW2EHDcE7iaY1SE0TTssmwK7JDa8dteK5XHcrLmbxw+3xWOFDJABMkAGyAAZIAPLwsAyy2sNfNUjCAKYGQmp7rmJ/0fDiQzl5r+H4I9MY2JuHpHoOMprNIn0h9EUjaP6inaxeWUEgWgM5SpL2o/qaBxN1/tRHIxjYm4a/hptWRMcQ17jKE8DUmFnDIHoCPbL99Y+W3pQGZnDxNwkStV6LdjVOY722DxmZucQCA1hl+oS3YXyoTgiysnnEXSWN7HPbOJp1qWzF8XBSQSnFzARi6Hykin5qlxzKpuN6enM5f3SiFttq56hNcoZCIbwX4kyOWLlJq6Wx1Jgl+UgpEg6GEzHJj8nf2SADJABMkAGyAAZ2NAMLL+8Hq7BngHpsjuP2voWFI8s2DKxm+vHMSFyNCsyFkcgLt8voL3VFLY643tbt9rrcWN7dXLRa3bFXViAIXQxVC5CXm1ya+4TCwuYmYwjMDaK4kQ9FhCJjKIyOImIVCs+ih2H75a8mnWZX8BMPI5AdBoTKhxxlFdZUu0ir67l7Te6Zo9HEjcJfP5RRABEggF36DOJKwXWPWY8gTIuZIAMkAEyQAbIABkgA2RgRRhYEXn1lfWiVsZPzs6rrGGk35KnNlSOy1DTGIoSkzoF4Bc3VVJYA1+u8hobQUEGaKwxrzNzC9Bf7R1G1sdNXmfCA0a2WG3XLNd4BPeZ+9l8KYTKtp7E/3PtNqyk2FYOK7NryqtWl80txrjhpLy7ZG/NGNnLW4NEnczYFqru2nPwu43rdYqrCLQlrFjAjD5kmBnYFTk4mZFlRpYMkAEyQAbIABkgA2SADCQZWBl5PVwDa/ImNUlTQlQNERPp0hvFkKxpVEu2MUd5ncgyBtQQuQVEJLurvazuyAnREzF122ci8wpMxCZR2zOE/Y4xsjnLq+ryq5fD7K5sZpFtdUkpS3p5TQqu2cAXjExr4LpksbuMmwJ6JlaTfaP+pq3OT6P6UkSbXCqOyksjCGoCOxMJ2dpMbz/+nTzAGAvGggyQATJABsgAGSADZIAM3B0GVkxere69NjFzkzUR3S6ZLWgetdL9N0XeauBz6TZs325qcGxyqkmbBZLte7d9qnWMMa+ByXnMSHdeGcsbDScyvrnKa/qy5iKmuSxj1d/MFo+FsbnKGHscDOhjYK3larDZklMlriK7xn6MWhpjhZPLzMFfb3brdomlFVO+J+PLWDAWZIAMkAEyQAbIABkgA2RgaQyssrya3Ylt2cBGlI7IsNdJNc7UktdgV1KWCnplZqR51GpjXtMLoREgm5y6CJft+wzyuuW0VY5Gs1uuVY4aY2blXGYbTpslzkVMc1kmCYWRxY6julduCMRRnsh6J5dJHEQVHdpzd1PlVS1nW8ZlGy6xTWyf3zFbTQbIABkgA2SADJABMkAGyMAiGVhlea2B9dicYG8/Cqo7sKstpiZwSozfLBtCu2Q556fhv9qD/Z2TiKispyWNLjLnEgybnGb73lVezQmQpsdRJAJbEUC5mnwq+WzYYiXd06j+ukuTQF3wjLLOjI2isitiexVdkOVc6pJSlpARj9gI9vs7jPG2Kcto+6yKqNmeVQZ1ZEgbw6st4xIPqyx65pUSmi1m/J6MkAEyQAbIABkgA2SADJCB5WJg1eXVd7gDxUOSSU3+m4gMoVDLEBa0TRqz7soic3GU96bONrwSmdfN9aO2cZ9YmEd7W0fyzkldclxopL8r+XlCDvVsZrK+8pcxXn0mar0AAA27SURBVDUXea3Bnt5pczIlc/KlTPJ62JwUC0B7m5U1zuWA0sua/hFDywUmt5tLG3EZckIGyAAZIANkgAyQATKwcRhYQXnNEtSyNjzmv44tmrR6FcT7qrtQeLYtjyxmlron5HY5ljO7YcuMznntp9+YIVrMenbceP5tXusvR124Ta8eEywX2SQDZIAMkAEyQAbIABlYbga8I68UI5dM7dIOgMKmMCqDcdUNO1tmerlB4/aX1paMH+NHBsgAGSADZIAMkAEysNEZoLyuY2kuDcszbecRGbJ3w97o0LP+PPGTATJABsgAGSADZIAMkIG1xwDldR3LKw/ItXdAss3YZmSADJABMkAGyAAZIANkwJ0Byivl9a53V+bB5n6wMS6MCxkgA2SADJABMkAGyAAZWDwDlFfKK+WVDJABMkAGyAAZIANkgAyQATLgeQYor4TU85Dy7tTi704xdowdGSADZIAMkAEyQAbIwHphgPJKeaW8kgEyQAbIABkgA2SADJABMkAGPM8A5ZWQeh7S9XKniPXgXU8yQAbIABkgA2SADJABMrB4BiivlFfKKxkgA2SADJABMkAGyAAZIANkwPMMUF4Jqech5d2pxd+dYuwYOzJABsgAGSADZIAMkIH1wgDllfJKeSUDZIAMkAEyQAbIABkgA2SADHieAcorIfU8pOvlThHrwbueZIAMkAEyQAbIABkgA2Rg8QxQXimvlFcyQAbIABkgA2SADJABMkAGyIDnGaC8ElLPQ8q7U4u/O8XYMXZkgAyQATJABsgAGSAD64UByivllfJKBsgAGSADZIAMkAEyQAbIABnwPAOUV0LqeUjXy50i1oN3PckAGSADZIAMkAEyQAbIwOIZoLxSXimvZIAMkAEyQAbIABkgA2SADJABzzOwrPJ6X3UXHjvd6AhCCwr8XSiocBh3RQcK/R24Lxs0ZW14zH8dW8oc62dbb919nyaOS6jn5tPXUXi2DZuXsI3Md5IaseWsS9vntT+pt1fa31Efsuk41jf6Mcr6Zz4fMD6MDxkgA2SADJABMpAfA8sqr3tC80BsBFt0OWmKYQbARKjXdqFb2D8HxEdRqC/r9nfdOCYwj9q6/Cq6emC0YU9HBEUXllbeLfVDqLwa0GIWRgBA4PrStqvHRbXX7Dj2uMX9rnzWi9rZpZZZ6u2V9nfUZ82xeffY0Tni34wrGSADZIAMkAEyQAbIwHIwsKzy6lOiGke5Jj5KkACHqDaicjxVaF0rvOYEwSE4Wixc65fmexW3sTDl9TDlNR9uuCx/OMgAGSADZIAMkAEyQAbWCwPLK6+HQ2hfWEBTkwVMF/xxIDIyiQlMo7rK+lyWA9pbzf9X9KByKI7I7DwiYzFU1rQkpS1FXhtR2DqK9tg8Zqan0R4ModDWpbgFu7piCEzOY2IyjqauHlvX5P29cTR19mBX16Sxv+goiqtrUHBlBE0xYx1/a5e9K22m8tnksx/V0WlMLAAzk3EEhsLJzHLO26iBlDE4DWBuDoFoDOV+iZOZee2SWE1jYm4eEbPsOpyb/x6CP2J8H4yMovjvzm7cVhvUIDXzmjl2PrMOwekFTMQmUdvRY49TWReKemMwvo+h8pKLyGeIQ2FnDIHeAdV+so32Dqve86itT9bbtV4Ztqvik6ns/jCaoiMouhRRDMxEhgz+stUnhc1kbPU24d+MCxkgA2SADJABMkAGyAAZyJ+BZZbXNlTHtIxqmZU1G0DTHBDoNAtsZWhFOqvCCMwDE9FRVHZFUD00jZmFeTQ1mdLlEIQ9wTlgYR6BYASVXSNojwMYj6BASWQbyqMLwGwc/i75fhSBWWAmGk50ZS4fA2Zm5xCMjBjriyROzyEYi6Fa7X8OM3o31Wzls8lrAEWyzzkgEo6gsqPX2K9jG5XBOCawgEBnW1LSte3suBqBf2wBiMdQ2TWEPUr6DXmdmZ9L1L1pcgGYi2G/tW7dKCILC4gMSd0i8EfngYVpVFa7g2KXV0fsemIISrsMDRiCWhZC+7xIeQzVbUMo7ZEbEnodOlAuZbZi3xNTNwcmpN2trs6OODjbWpVnbgGR+KRqP6PrtdR7ARPT8/Z6zcdRbtXLsd2U+GYru2JsATPxObQLV6q7dg71UevZexrwpOTOGuPCuJABMkAGyAAZIANkgAzky8Ayy2sN1FhWa9xr66SSq6LDNSgeAWbCA0rW9GV2iIzGRkz5NBpUfTYeMcRPl9eyISNj26ZlE6tCqI2Oo/TzGvjUsnPw6+NNq0cQxAKarhjbFnm1jcutlzG1c/Cr7KYs04PaaSDYY4hl1vJZ4ph4T802KimLj2JHYpkaFPRM28VT+04aVa3j0m040tuRFN4qqZs1HrQRpWaMkxMwmZ8N9SfX0fZjk9cLo4hIdtwSQlnO/KxS5LmsDTuuDmCXluUuCi8AUbNr8xUZ22xff7M53tmS12yxNOR1EsXaPqyMs7CTrFeHukkyY9Yra3yzld2U1/YWjasc6mPwRnnN9yTE5fnDRQbIABkgA2SADJABMpALA8sur4kL+rIaKLkZCSlx2nw9DqjJgYzsbKS/C77DZqY2amQKJVuoXuG5hPQa2zMFTWRoYRLFmoDpld4iQjg9jl22782uy2p/NVDyqkuhS/ZMljEmmMqhfLZ9CYROeTW3MdBjF0i/yKIuzXaA08mrJYJGva3MtqxrZLeDQ2YMzViqDK51M8FRVl1eVexm4yr7nGiHrhgiIv6JbuAtKKgZQHnXCGojcQRnAZixVOunTMBlZtxV5jV7LPXyJNvVyDgnupibdUhO+JVrfNOX3caYuf3s9bFullBek21lZ5ifMx5kgAyQATJABsgAGSADS2Fg+eXVlKj2VkMaA9fNbJbKEsq4V124jCynGh8ajSNge40Y3WH1zGunJcDuEOwamFddiJMZOlnOPjlUfvKaQ/kcQpgqr2Ymt0vL6sk6ZREta5pan/zl1ZC8iXFnHB1jb7Xy6rKoYqfG2KauXy1Z62qje7cxDncStT1hVEbmE/Kq1k+RZF3ks8dSL08ScpOXenuMNndNmzdDcohvlrK7yWv2+lBek21kbxt+zniQATJABsgAGSADZIAM3A0GVkBejS7CE+EYgrZJmkyZDU9iZi4G6UosFSqPJrsTu1ZQl1eVrdQnfpJtaM8BlW7KKZlZ+3jb/OQ1h/JpMmiUXxc2o47SZRpmBjpRR2vcb8r6xjr5y6spcWZ358R+0mxfvrfJotwY0NrFub6V6dS7PqvMupXFlvWdsVc3LJJjXrO1ta08iXIbUm5147bKpXdZzhbfrGXXGbP2m0N9pCv1Y7k8q9jaJt/tvQ8YD8aDDJABMkAGyAAZIANkIAMDKyKvqotwyuNxTFmSzzWRM8ZFzqP2ijnDcFkXKmXin2gY/yUVsYmFOdZxLGLOMNyIQsnALUyiVMZJlhmiOhHqN2cYbsF+efbsfHIcZb7ymrV8KcE25DXSm5yxOGUbFf1qXO3MSEgbx2lIqyVnSuQmR7SZlA2JS99t2BxHOx9HpTXDsLmfiQH7M3Zt+7Ce85oSO4ltHDMLcZR/VAPVpnOTKK4wyrn50oia0MnqNuwzJ0VKxN5qR+3ZtClxsJYx2zqTvEKr131XZJzyAtpbjWx2ynYd8c1adhtjZjvkUB9fXQSBsRhK9XHCKTzY29WKPd8ZFzJABsgAGSADZIAMkAEykJmBFZFXn5lxM8aNagVKTOCjd6FtxK6uuHq8DBbEbGX238mkEDjFojqEJplhWP7J8gtzaGpJTmK0+VJEzTBsLgHMxVF9Kbm/fOVVuh1nLJ+LrBRcj2NGlc8an5u6jZmxEdvkRyngWl1dYT1SKLu8+g53oDg0be7biMBMbBR7bBMgJdvDKYu22KnYzqO9zYxtWS/8MaOBZuYBkclarduwlH9z/aghtGY7RgYiqJ1NZl6zxdJZHiMmZrfhrnFEVJmMhg/292rinyW+2cruZMxs02z12RKQqaoX0N6SjGlKO7rwwWUYLzJABsgAGSADZIAMkAEykJ2BlZHXRV2wS/ffLjx2OimamRp08+nrKDzbpgmMvfL3VXehsFp7XuyiyqRvM7/yuZfd2EaBmb10X0bf5yL/Vt1Zu7DY/ajYpYlttriLoG45m60dFxtLWe86tqSRcaMLefp6Zy+7W7xzqY/bevxs2fhe8rHMtmHbkAEyQAbIABkgA2RgLTDgYXklQGsBIJaRnJIBMkAGyAAZIANkgAyQATKwEgxQXpm14aBwMkAGyAAZIANkgAyQATJABsiA5xmgvBJSz0O6EndxuA/eLSQDZIAMkAEyQAbIABkgA95mgPJKeaW8kgEyQAbIABkgA2SADJABMkAGPM8A5ZWQeh5S3gHz9h0wtg/bhwyQATJABsgAGSADZGAlGKC8Ul4pr2SADJABMkAGyAAZIANkgAyQAc8zQHklpJ6HdCXu4nAfvFtIBsgAGSADZIAMkAEyQAa8zQDllfJKeSUDZIAMkAEyQAbIABkgA2SADHieAcorIfU8pLwD5u07YGwftg8ZIANkgAyQATJABsjASjBAeaW8Ul7JABkgA2SADJABMkAGyAAZIAOeZ4DySkg9D+lK3MXhPni3kAyQATJABsgAGSADZIAMeJsByivllfJKBsgAGSADZIAMkAEyQAbIABnwPAOUV0LqeUh5B8zbd8DYPmwfMkAGyAAZIANkgAyQgZVgQOT1/wMqyfi0cs7eLgAAAABJRU5ErkJggg=="
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Improvement"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import packages\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import GradientBoostingRegressor, AdaBoostRegressor\n",
    "from sklearn.ensemble import AdaBoostClassifier, GradientBoostingClassifier\n",
    "import xgboost as xgb\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "from sklearn.metrics import accuracy_score, roc_auc_score, confusion_matrix\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "SEED=42"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3071: DtypeWarning: Columns (19,20) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    }
   ],
   "source": [
    "# Load train data\n",
    "mailout_train = pd.read_csv('Udacity_MAILOUT_052018_TRAIN_reduce_mem.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>...</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>RESPONSE</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1763</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1771</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>1776</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>...</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>1460</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>...</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>1783</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 368 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0   LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  ALTER_KIND2  \\\n",
       "0           0  1763         2         1.0       8.0          NaN          NaN   \n",
       "1           1  1771         1         4.0      13.0          NaN          NaN   \n",
       "2           2  1776         1         1.0       9.0          NaN          NaN   \n",
       "3           3  1460         2         1.0       6.0          NaN          NaN   \n",
       "4           4  1783         2         1.0       9.0          NaN          NaN   \n",
       "\n",
       "   ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ...  VK_DHT4A  VK_DISTANZ  \\\n",
       "0          NaN          NaN                   8.0  ...       5.0         2.0   \n",
       "1          NaN          NaN                  13.0  ...       1.0         2.0   \n",
       "2          NaN          NaN                   7.0  ...       6.0         4.0   \n",
       "3          NaN          NaN                   6.0  ...       8.0        11.0   \n",
       "4          NaN          NaN                   9.0  ...       2.0         2.0   \n",
       "\n",
       "   VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008  WOHNLAGE  ZABEOTYP  RESPONSE  \\\n",
       "0      1.0             6.0             9.0       3.0         3         0   \n",
       "1      1.0             4.0             9.0       7.0         1         0   \n",
       "2      2.0             NaN             9.0       2.0         3         0   \n",
       "3     11.0             6.0             9.0       1.0         3         0   \n",
       "4      1.0             6.0             9.0       3.0         3         0   \n",
       "\n",
       "  ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0         2                    4  \n",
       "1         2                    3  \n",
       "2         1                    4  \n",
       "3         2                    4  \n",
       "4         1                    3  \n",
       "\n",
       "[5 rows x 368 columns]"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mailout_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "mailout_train = mailout_train.drop(['Unnamed: 0'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LNR</th>\n",
       "      <th>AGER_TYP</th>\n",
       "      <th>AKT_DAT_KL</th>\n",
       "      <th>ALTER_HH</th>\n",
       "      <th>ALTER_KIND1</th>\n",
       "      <th>ALTER_KIND2</th>\n",
       "      <th>ALTER_KIND3</th>\n",
       "      <th>ALTER_KIND4</th>\n",
       "      <th>ALTERSKATEGORIE_FEIN</th>\n",
       "      <th>ANZ_HAUSHALTE_AKTIV</th>\n",
       "      <th>...</th>\n",
       "      <th>VK_DHT4A</th>\n",
       "      <th>VK_DISTANZ</th>\n",
       "      <th>VK_ZG11</th>\n",
       "      <th>W_KEIT_KIND_HH</th>\n",
       "      <th>WOHNDAUER_2008</th>\n",
       "      <th>WOHNLAGE</th>\n",
       "      <th>ZABEOTYP</th>\n",
       "      <th>RESPONSE</th>\n",
       "      <th>ANREDE_KZ</th>\n",
       "      <th>ALTERSKATEGORIE_GROB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1763</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1771</td>\n",
       "      <td>1</td>\n",
       "      <td>4.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>13.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1776</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1460</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>...</td>\n",
       "      <td>8.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1783</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.0</td>\n",
       "      <td>53.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 367 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    LNR  AGER_TYP  AKT_DAT_KL  ALTER_HH  ALTER_KIND1  ALTER_KIND2  \\\n",
       "0  1763         2         1.0       8.0          NaN          NaN   \n",
       "1  1771         1         4.0      13.0          NaN          NaN   \n",
       "2  1776         1         1.0       9.0          NaN          NaN   \n",
       "3  1460         2         1.0       6.0          NaN          NaN   \n",
       "4  1783         2         1.0       9.0          NaN          NaN   \n",
       "\n",
       "   ALTER_KIND3  ALTER_KIND4  ALTERSKATEGORIE_FEIN  ANZ_HAUSHALTE_AKTIV  ...  \\\n",
       "0          NaN          NaN                   8.0                 15.0  ...   \n",
       "1          NaN          NaN                  13.0                  1.0  ...   \n",
       "2          NaN          NaN                   7.0                  0.0  ...   \n",
       "3          NaN          NaN                   6.0                  4.0  ...   \n",
       "4          NaN          NaN                   9.0                 53.0  ...   \n",
       "\n",
       "   VK_DHT4A  VK_DISTANZ  VK_ZG11  W_KEIT_KIND_HH  WOHNDAUER_2008  WOHNLAGE  \\\n",
       "0       5.0         2.0      1.0             6.0             9.0       3.0   \n",
       "1       1.0         2.0      1.0             4.0             9.0       7.0   \n",
       "2       6.0         4.0      2.0             NaN             9.0       2.0   \n",
       "3       8.0        11.0     11.0             6.0             9.0       1.0   \n",
       "4       2.0         2.0      1.0             6.0             9.0       3.0   \n",
       "\n",
       "   ZABEOTYP RESPONSE ANREDE_KZ ALTERSKATEGORIE_GROB  \n",
       "0         3        0         2                    4  \n",
       "1         1        0         2                    3  \n",
       "2         3        0         1                    4  \n",
       "3         3        0         2                    4  \n",
       "4         3        0         1                    3  \n",
       "\n",
       "[5 rows x 367 columns]"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mailout_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define the response variable\n",
    "X = mailout_train.drop('RESPONSE',axis=1)\n",
    "y = mailout_train['RESPONSE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape before (42962, 366)\n",
      "{'OST_WEST_KZ', 'D19_LETZTER_KAUF_BRANCHE', 'CAMEO_INTL_2015', 'CAMEO_DEU_2015', 'CAMEO_DEUG_2015', 'EINGEFUEGT_AM'}\n",
      "shape after corr (42962, 276)\n",
      "Number of nan Values 0\n",
      "shape after (42962, 321)\n"
     ]
    }
   ],
   "source": [
    "# Data Pre_processing \n",
    "df_mailout_train  = data_cleaning(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the data\n",
    "X_train,X_test,y_train,y_test = train_test_split(df_mailout_train,y,stratify=y,random_state=SEED,test_size=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Compare models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(df_mailout_train, y, stratify=y, test_size=0.2, random_state=SEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_and_predict(model, X_train, y_train, X_test, y_test):\n",
    "    \"\"\"\n",
    "    Fit a model on X_train, y_train\n",
    "    predicts on X_text, y_test \n",
    "    Calculate AUROC on predictions made on test data\n",
    "    \n",
    "    Outputs - AUROC score, time elapse for training and prediction    \n",
    "    \"\"\"\n",
    "    start = time.time()\n",
    "    model = model.fit(X_train, y_train)\n",
    "    \n",
    "    roc_score = roc_auc_score(y_test, model.predict_proba(X_test)[:,1])\n",
    "    \n",
    "    end = time.time()\n",
    "    time_elapsed = end - start\n",
    "    \n",
    "    return roc_score, time_elapsed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "models = [(\"LogisticRegression\", LogisticRegression(random_state=SEED)),\n",
    "         (\"DecisionTreeClassifier\", DecisionTreeClassifier(random_state=SEED)),\n",
    "         (\"RandomForestClassifier\", RandomForestClassifier(random_state=SEED)),\n",
    "         (\"GradientBoostingClassifier\", GradientBoostingClassifier(random_state=SEED)),\n",
    "         (\"AdaBoostClassifier\", AdaBoostClassifier(random_state=SEED)),\n",
    "         (\"XGBClassifier\",xgb.XGBClassifier(random_state=SEED))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:888: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[19:30:13] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.\n"
     ]
    }
   ],
   "source": [
    "results = {\"Model\":[],\n",
    "          \"AUCROC_score\":[],\n",
    "          \"Time_in_sec\":[]}\n",
    "\n",
    "for name, model in models:\n",
    "    roc, time_ = train_and_predict(model, X_train, y_train, X_val, y_val)\n",
    "    results[\"Model\"].append(name)\n",
    "    results[\"AUCROC_score\"].append(roc)\n",
    "    results[\"Time_in_sec\"].append(time_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>AUCROC_score</th>\n",
       "      <th>Time_in_sec</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LogisticRegression</td>\n",
       "      <td>0.689629</td>\n",
       "      <td>1.935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DecisionTreeClassifier</td>\n",
       "      <td>0.540728</td>\n",
       "      <td>2.31003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RandomForestClassifier</td>\n",
       "      <td>0.578212</td>\n",
       "      <td>13.6376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>GradientBoostingClassifier</td>\n",
       "      <td>0.76548</td>\n",
       "      <td>53.774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AdaBoostClassifier</td>\n",
       "      <td>0.734101</td>\n",
       "      <td>14.3908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>XGBClassifier</td>\n",
       "      <td>0.695248</td>\n",
       "      <td>16.8857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        Model AUCROC_score Time_in_sec\n",
       "0          LogisticRegression     0.689629       1.935\n",
       "1      DecisionTreeClassifier     0.540728     2.31003\n",
       "2      RandomForestClassifier     0.578212     13.6376\n",
       "3  GradientBoostingClassifier      0.76548      53.774\n",
       "4          AdaBoostClassifier     0.734101     14.3908\n",
       "5               XGBClassifier     0.695248     16.8857"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = pd.DataFrame.from_dict(results, orient='index').transpose()\n",
    "results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Grid Search "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### AdaBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Define the paramaters for GridSearch\n",
    "param_grid = {\"n_estimators\": [50,100,150,200],\n",
    "              \"learning_rate\": [0.01,0.1,0.5,0.9,1.],\n",
    "              \"algorithm\":[\"SAMME.R\"]\n",
    "              }\n",
    "\n",
    "\n",
    "adaboost_grid = GridSearchCV(estimator = AdaBoostClassifier(random_state=SEED), \n",
    "                           param_grid = param_grid, \n",
    "                           scoring = \"roc_auc\", \n",
    "                           cv = 5, n_jobs = -1, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Score:  0.7597514028459369\n",
      "Best Params:  {'learning_rate': 0.01, 'max_depth': 2, 'min_samples_split': 3, 'n_estimators': 1000}\n",
      "Wall time: 1h 46min 59s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "adaboost_grid.fit(X_train, y_train)\n",
    "\n",
    "best_adaboost = adaboost_grid.best_estimator_\n",
    "\n",
    "print(\"Best Score: \", adaboost_grid.best_score_)\n",
    "print(\"Best Params: \", adaboost_grid.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC score on validation data: 0.7613\n"
     ]
    }
   ],
   "source": [
    "preds_adaboost = best_adaboost.predict_proba(X_val)[:,1]\n",
    "\n",
    "print(\"ROC score on validation data: {:.4f}\".format(roc_auc_score(y_val, preds_adaboost)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def plot_feature_importances(model, model_name, feature_names, num_features=10):\n",
    "    \"\"\"\n",
    "    Plots top n feature importances for the given model\n",
    "    \"\"\"\n",
    "\n",
    "    \n",
    "    feature_importance_values= np.zeros((len(model.feature_importances_)))\n",
    "    \n",
    "    feature_importance_values += model.feature_importances_\n",
    "\n",
    "    feature_importances = pd.DataFrame({'feature': feature_names, 'importance': feature_importance_values})\n",
    "\n",
    "    # sort based on importance\n",
    "    feature_importances = feature_importances.sort_values('importance', ascending = False).reset_index(drop = True)\n",
    "\n",
    "    # normalize the feature importances to add up to one\n",
    "    feature_importances['normalized_importance'] = feature_importances['importance'] / feature_importances['importance'].sum()\n",
    "    feature_importances['cumulative_importance'] = np.cumsum(feature_importances['normalized_importance'])\n",
    "    \n",
    "    fig = plt.figure(figsize=(8, 5))\n",
    "    ax = fig.add_subplot(111)\n",
    "    \n",
    "    ax.barh(list(reversed(list(feature_importances.index[:num_features]))), \n",
    "                feature_importances['normalized_importance'][:num_features], \n",
    "                align = 'center')\n",
    "\n",
    "    # Set ticks and labels\n",
    "    ax.set_yticks(list(reversed(list(feature_importances.index[:num_features]))))\n",
    "    ax.set_yticklabels(feature_importances['feature'][:num_features])\n",
    "    ax.set_xlabel('Normalized Importance')\n",
    "    ax.set_title(f'Feature Importances ({model_name})')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_feature_importances(model=best_adaboost, model_name=\"AdaBoost\", feature_names=df_mailout_train.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_param_grid = {\"max_depth\": [5,10,20,30],\n",
    "              \"learning_rate\": [0.01,0.1,0.5,0.9,1.],\n",
    "              \"gamma\":[0.1,0.5,1.0],\n",
    "              \"n_estimators\":[50,100,150,200]\n",
    "              }\n",
    "\n",
    "xgb_grid = GridSearchCV(estimator = xgb.XGBClassifier(objective=\"binary:logistic\", \n",
    "                                                            n_jobs=-1, eval_metric=\"auc\",\n",
    "                                                            silent=1,random_state=SEED), \n",
    "                           param_grid = xgb_param_grid, \n",
    "                           scoring = \"roc_auc\",\n",
    "                           cv = 5, n_jobs = -1, verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 240 candidates, totalling 1200 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  33 tasks      | elapsed:  8.3min\n",
      "[Parallel(n_jobs=-1)]: Done 154 tasks      | elapsed: 78.4min\n",
      "[Parallel(n_jobs=-1)]: Done 357 tasks      | elapsed: 168.6min\n",
      "[Parallel(n_jobs=-1)]: Done 640 tasks      | elapsed: 294.5min\n",
      "[Parallel(n_jobs=-1)]: Done 1005 tasks      | elapsed: 452.8min\n",
      "[Parallel(n_jobs=-1)]: Done 1200 out of 1200 | elapsed: 551.4min finished\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:888: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[05:15:42] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:541: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "Wall time: 9h 11min 34s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "xgb_grid.fit(X_train, y_train)\n",
    "\n",
    "best_xgb = xgb_grid.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Score:  0.7633140979328148\n",
      "Best Params:  {'gamma': 0.1, 'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 50}\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Score: \", xgb_grid.best_score_)\n",
    "print(\"Best Params: \", xgb_grid.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC score on validation data: 0.7703\n"
     ]
    }
   ],
   "source": [
    "preds_xgb = best_xgb.predict_proba(X_val)[:,1]\n",
    "\n",
    "print(\"ROC score on validation data: {:.4f}\".format(roc_auc_score(y_val, preds_xgb)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_feature_importances(model=best_xgb, model_name=\"XGBClassifier\", feature_names=df_mailout_train.columns)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kaggle Submisson"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AdaBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(learning_rate=0.01, max_depth=2, min_samples_split=3,\n",
       "                           n_estimators=1000)"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fit model\n",
    "best_adaboost.fit(df_mailout_train, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make prediction on test data\n",
    "preds_test_adaboost = best_adaboost.predict_proba(mailout_test_clean)[:,1]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert to dataFrame with appropriate columns\n",
    "kaggle_adaboost = pd.DataFrame(index=mailout_test[\"LNR\"], data=preds_test_adaboost)\n",
    "kaggle_adaboost.rename(columns={0: \"RESPONSE\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model to csv for submission\n",
    "kaggle_adaboost.to_csv('submission5.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\xgboost\\sklearn.py:888: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].\n",
      "  warnings.warn(label_encoder_deprecation_msg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[05:55:55] WARNING: C:/Users/Administrator/workspace/xgboost-win64_release_1.3.0/src/learner.cc:541: \n",
      "Parameters: { silent } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, eval_metric='auc',\n",
       "              gamma=0.1, gpu_id=-1, importance_type='gain',\n",
       "              interaction_constraints='', learning_rate=0.1, max_delta_step=0,\n",
       "              max_depth=5, min_child_weight=1, missing=nan,\n",
       "              monotone_constraints='()', n_estimators=50, n_jobs=-1,\n",
       "              num_parallel_tree=1, random_state=42, reg_alpha=0, reg_lambda=1,\n",
       "              scale_pos_weight=1, silent=1, subsample=1, tree_method='exact',\n",
       "              validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fit XGBoost model\n",
    "best_xgb.fit(df_mailout_train, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make prediction\n",
    "preds_test_xgb = best_xgb.predict_proba(mailout_test_clean)[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "## convert to dataFrame with appropriate columns\n",
    "kaggle_xgb = pd.DataFrame(index=mailout_test[\"LNR\"], data=preds_test_xgb)\n",
    "kaggle_xgb.rename(columns={0: \"RESPONSE\"}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model to csv for submission\n",
    "kaggle_xgb.to_csv(\"submission3.csv\")"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5gAAADACAYAAABh/cy6AAAgAElEQVR4Ae29629TZ/63y19R9d3skbZGGmlrpHnRF93Poz4S0ghpcN90/ySkaf3mSavZFd2VpqBdCQYN6oFdhXTa0NJwCvPj0EL6EAi/QANpSE3SEEJIXIeQmJBDA8nEnqQO9RiCgeizda+D17LjnMiiJOFCcn1Ya93rXte6nPrj730vrxH/IAABCEAAAhCAAAQgAAEIQAACARBYE0AbNAEBCEAAAhCAAAQgAAEIQAACEBABEwkgAAEIQAACEIAABCAAAQhAIBACBMxAMNIIBCAAAQhAAAIQgAAEIAABCBAwcQACEIAABCAAAQhAAAIQgAAEAiFAwAwEI41AAAIQgAAEIAABCEAAAhCAAAETByAAAQhAAAIQgAAEIAABCEAgEAIEzEAw0ggEIAABCEAAAhCAAAQgAAEIrElNToobDHAAB3AAB3AAB3AAB3AAB3AAB5bqABVMvmSAAAQgAAEIQAACEIAABCAAgUAIEDADwUgjEIAABCAAAQhAAAIQgAAEIEDAxAEIQAACEIAABCAAAQhAAAIQCIQAATMQjDQCAQhAAAIQgAAEIAABCEAAAgRMHIAABCAAAQhAAAIQgAAEIACBQAgQMAPBSCMQgAAEIAABCEAAAhCAAAQgQMDEAQhAAAIQgAAEIAABCEAAAhAIhAABMxCMNAIBCEAAAhCAAAQgAIHVT2Bo+JY+3b1P23fsfKyb2da0wb/VS4CAuXrPLUcGAQhAAAIQgAAEIACBQAl8snvvYwVLfyA1IZN/q5cAAXP1nluODAIQgAAEIAABCEAAAoEScIPi4za61O0fd79s98sRIGD+cqzZEwQgAAEIQAACEIAABFY0gaUGxKVuv6LhPSOdJ2A+Iyeaw4QABCAAAQhAAAIQgMBSCSw1IC51+6X2n+2fPAEC5pNnzB4gAAEIQAACEIAABCCwKggsNSAudftVAXGVHwQBc5WfYA4PAhCAAAQgAAEIQAACQRFYakBc6vZBHQftPDkCBMwnx5aWIQABCEAAAhCAAAQgsKoILDUgLnX7VQVzlR7MMgqYDzV1J63J+9PBoZ6+o0vnojo7lF1ym1P9Pfr83JAml9xSYQMPNZUc06WuMY2ll97PwtZ5DgEIQAACEIAABFYjgWwmoYHYgBKZ+T8/ZTMZZYreimybzSgxENNAMqMiS22Uj7LKJAcUG0io2O4Xur+i600t77O11IC41O2XNx16Zwgso4A5pA9KG/RKQyq4MzPeoz+VNujF2tElt9lZ3aA1Za06e2fJTXkNjPdp62cNWlPq3dYe738CIdbbJY8gAAEIQAACEIDAyiaQUPv+LQqHQgpZt7C2HI0pM+tBpRT5yF238L5Kvb7tEm0V2pxrN6TQG+VqSfpWkJQdOKMdb/jaCW9RZVvCt1Kvqvxt+B9/FFHuk242pkr/MvfxMX+PfM0uk4dLDYhL3X6ZYKAbcxBY3QHTHHgmo6kgiqLTWU1mHs6BcrGL0qo+0KA1e6PquWM6OK3JaLteLG3Q1vZZvy9b7E5YHwIQgAAEIAABCKwqAolvtysU2qEzP9qRMnO9ygqFOy7motuM451ZKRzQmfdCCn/R7gXTgZPaHApr57lb9muPUmo/tFmhTSc1kGtxQCffDuntg+1KPbJfTHy7Q6HQdtX/M7eSZuxvvF2V/3dIm2u8ljQe0Y7QFp3pK6iuUsH0QPJoRRIINmBmU7rUENWmAxG98p8d+rLbK/dNdnWprDquQR+mwUsdKjs37FTs3ApmUoOtHdpY0aSNX3fpQtILdW4bPeNDOnC0xdpHdf89yRoK26aSz1q0yQxjzQXKpM5Wd+hAV9rZ67Qmh/t04OtWvWKt26dB/9ddi+r/tCb7evT50RatrWhVWcOwxny50D22QTO01qxzoE0Hoh4PKa2ejn51jvo2UlIHdjfoj+cKvirzMVM2pQvn2rSxIqKSo106O+w/gIca6+rO65PN4qF6mjpUVtuvMX9bw3GVVUd1YY7d+VfnMQQgAAEIQAACEHi6BJyAV+0LapJ6j4UV+mu9/HXEufqZjVfp7dBmnfQ1M1DztkJv+8OkZIfAsA7HnM9rVigMqSrua73Ya77F5mHq4g6FwhVq/9m3IF6lUGinWmbPxb6Vl8/DpVYgl7r98iFBT2YjEGDATKl6f4Oe/6xFnzf1q7q2xarGbWy2w91YQ5PWlLbrkq8nl742FbxuJ/TYAfP35RH98WCbyqpNYDTDUpv05aidGO02GvXi3hZt9S0v2R9RydcdKjvapF+bIbGnbzt7cUOr/c6d6rYrhK983a2zrd3aWtGgNeUd6rSaX0z/pzXY0KTnzZDeox12X8vNsUedtiTr2D5t1Nrdpq8d2rTXDINt1AfuHygfB/fhVLxDf5yrgpkZ0gfWfuw2tx6M6HmrzXtWE2ORJj1fFlFJbVwXmqIqMesejFsBfqq9Tc+XRnRg2N2b1Hm6UWs+dY/fe51HEIAABCAAAQhAYFkScMLc4ev5vctGKxcR1uwhs3nVS0kDXxcJqZl2VYRC8qqjCdX/LaS3fUNyExd3KlwYHv3dy/aq6u2C6qUZamv1+bBa+iI6U31SJ6vr1f5Pf+HA38jyebzUgLjU7ZcPCXoyG4HgAuZPPSox4ajDrTiaCl+/Op0K5EID5poDPd4cxOnbKvu0Qc9/PWT1326jRdU/OYfj7HNtnVuXm9al4zNDqzuv05pHubvLq6JmkrrUlbQrnovp//1+bS1r0B9z+5XkbF/SZAdqK2CWteqC+3dielSfm8DsHIv/hHTWRrT20wat+XtEH7TO/jXW4LmINQ8016bu6cLXEa09akLkHXvI7Vf9XtPjo7oQd9qbHtIHps+56qjN9ve5MO5txiMIQAACEIAABCCwLAnMVi20qoE7FBmfv9fFqpdmq2zssMKhzaq67n54y6r32GZ7nqd/XmSmVyf/9qZCb2zWlo1hhTaVq37APyItvw9Fq5duVTMU1pullTpZfVKVH5UoZPYfn72t/JafzrOlBsSlbv90jpq9LoZAcAFTdgXQVBw31XbrbDypKd/7Y6EB0w2D7kFYobCi2wqFM9vIr1CabYpVRd023Qrm2oNtOtA0rJ6ffB1cTP+72+1qoJ17na6mdfZgg9Yc7ZMZOp/fD7NKSl+aKmaRgDk52K8LHXF9biqSZRF9ELMrki4D+97hezxvp3mrWBVMq6oa1ZettzWYN2d0Wp2nGrXGDdjxDv2+oKKZ1xhPIAABCEAAAhCAwHIjsOSAWbx6aR9mRjEz59KEvi1brPC4+WC9Tn4U0o5Gd/BtQi2fllih8sylmGKxFp00wXBTlXqLzZ2cpXqZw+rM47SfZxU7WKSKmlt5eTxYakBc6vbLgwK9mItAgAHTfPVj5mCa+ZON+m2ZGd7qhaWZ4bAwhNlh0R1S63Z6sC5iD+OUNLONxQVM60I6Zg6mmRNpKoalDfr1gR4NunM2F9r/WLvWlDbpS7dwanXWqZ7u77GG/C4mYLrHKt3T2cMNXgj0FswZUL3V7DmYZf/ZpLV/t49vbfWQFXitdXyh0hoe64ZNrwEeQQACEIAABCAAgeVLINWinaGQKqP+IoGUaatQKFShdrf4OMsR2NXL/AvyFK6aud2rWCym3tumsV4dDoVU0WY3bA9rLdzemRfqv4CP06hdvaxUrFj4LNxxbtjs/MdRZNNf7KWlBsSlbv+LHSg7emwCAQfMrHfF1ukJu2LnzLG0w2GbLrlhzg1ThXMwT7nzJ80xOVdadeYRBhEwp+67Q3ilqWibfl/aoA9iDr/sfP135pCOdeuVwrmSznBed8jpvAHzzrAOVHfobC7dmj44IdXMi5xxSqftqmhBKLSqn90TToh8qKnc74hOa7DezHv1B+ExHfjM/BRM3Bp67FZ2Z+yKFyAAAQhAAAIQgMCyJGDPgQwfivl+o9Kp/P1tvov8ONXLg/5tvYPMxOt1srpdt7yXZAfScrU4F+exAmOoUvmX1Ego8mFIIf8wWtOGU73cfs6tfvoaVla3mk/q5Le93lVsJVlXyA1X+a5a699meTz+/z7eJTckPu79J7v3Lo+DoRdPhEBwAXO4y7pAzZ/qx+yQmRpW2e4GPX+4z55T2WOGZDboxa/i6kkmdam2xa5yFgTM58uaVNaV0uSdlDobWue5UNBiKphZXfjKXNSnXRdSD6XprMYiLdZFcj430xbn6X9+uHWGw5a36ezte5rKpNRZZy76E9GBITtBzxswTSAtN/1pU/Wgfbw9TW1aW+ZdpGiqu8O62u0BN4TGOywef6q7rbE7aY3Fu/Sn3Pr2FWif39+tQRMys2ldOmXP2fT/dqc9j7NBz5f65rI+EbVoFAIQgAAEIAABCARPwFQrzVzJiosDSmVSGrhofrsynKsymmlJLV9s0Y5j+b+Nac+xLKw++vrnDL/dfKhFt1IZJeL1qtgU0uavB7ww+896bQ+F5K6TMfs/t1PhvP3bbdphcfbqZf5xZJSIVWl7eObFgHw9XBYPh4ZvyQTEpYRL0wb/Vi+B4AKmpjXY3GoFJDP01Bp+ujeqztxQhXvqrLWv8mqW/bF6SGeLXEX2g9YhfWAuhuO08Urt7dwQz/yQZ07KYgKm+U3M2/rcupqr035ZRJuaJ5yzO3f/Z+w7M6oDBxutK8lafS1r0ufduYNd2BzM8SGV7fe30ag/VXs/szLZ2mq1vzU3DGRak23t+qMZfuzwWXs47g3xHY1rk4+dmQ/7eU/BfE6n+pp3MaXV6zdHBgEIQAACEIDAqiOQ1cB/7VBJKGRfgCf8psq/veWFQN3SmU0hhT+M+H62xKl8zlK9dBFl+s5oxxtOu6ESbT/Y7mvDXsuss9Nc3Mfd/xvbVXmpoEo5FVNlOKTi1Ut3b1ndulipLWFvfzuq8yua7prcQ2AlEQgwYLqH/VBTd9KazA3VdF9fzP20ptLpvIsELWbrede9n9HknYLgldtokf03baXz5wHkmlrog+y9OfpTrJF5+phJL71PxXbLaxCAAAQgAAEIQGC5EHiUVSaT8QXL4DqWNe3mXYCnSNvZjDKZJX4GdJq19ldkF7wEgZVI4AkEzJWIgT5DAAIQgAAEIAABCEAAAhCAwFIJEDCXSpDtIQABCEAAAhCAAAQgAAEIQMAiQMBEBAhAAAIQgAAEIAABCEAAAhAIhAABMxCMNAIBCEAAAhCAAAQgAAEIQAACBEwcgAAEIAABCEAAAhCAAAQgAIFACBAwA8FIIxCAAAQgAAEIQAACEIAABCBAwMQBCEAAAhCAAAQgAAEIQAACEAiEAAEzEIw0AgEIQAACEIAABCAAAQhAAAIETByAAAQgAAEIQAACEIAABCAAgUAIEDADwUgjEIAABCAAAQhAAAIQgAAEIBBYwLz7Y6tqz9YW3Fo1fHcm5PFrter+18zXV+srQRzvWHOr1n4Wyb8djmtsqdDG49rktjMY1dqyVp29I8n/et4+butzfz8OtOnzplFNuevMup27QtD3pj9RdQbdLO1BAAIQgAAEIAABCEAAAosmEFjA1KOsslnfLTOo749c0eijmX1KdOxVZ3Lm66v1lSCOd6yhSWu+6tPknbR3S2eXjmysW6/s7baD6nRGg4Mpu03/63l7GdIHpU060Of0Izmm6q8a9UrDfNvlNRLgE9Ofdl0KsEWaggAEIAABCEAAAhCAAAQej0BwAbNg/+l4nep60wWv2k+twDV6V6PxmGI/dGvwp/yglP15WL0/xBS7dlMJXwU0m+xVb9Jb1//cPB6ezCph2vzR2W82rWFnHzf/5W1X2KnsT4NF92fWM8u6f4ip++a4slMJ9bptm4XTd5W42a3YDzH13kprtj3YAfORJofMut26mXQO6tGkBn+4qXF/CJ8cVmxgUv6XzK6sgPn1UGHXnefTmuyL60B1h8pqu3Up+TC33uClLl0aHVV1bYcOdDlckv32uuf6NNjvC5hK6mx1XIPWDv2v55qTZAfML/2l09FuvbK/xw6pTjDt6evR56Y/5/o1lgPja99qsuD5dEY9TVGVVUf1ZfcdTXZ16eywu+97Guzodtrs0+CdaWdBQcDMpnSpwbTRoQOtSU26q5m1TYBu7bKWfd4w7OuXuw/uIQABCEAAAhCAAAQgAIGlEHgyAfPRqK4cK169NJ01gav6XKduTqSVTvbq+1M16v7JPoxHyU7VnuvU8ORde1ntBd10ctHdm/Wqv+klTv9z87im9qI6h0Y1OmkSzbhitXXqHE0re3dcvY1f6UqRcurdgQv6ytnf3Ymb+v7U9xp2xntmf/w+tyyd7Fbrt3Wq7kg4vNO62Vir7+MJpe9Oajhap9qOxIxg6B5v3betit2a1N10Qr0ttbpgHdQjjbbl9yvRkf/c2dmcAXMs0qRfV7Tr7GBKY/FubSxv0VmH56WvG/Xi3lYdaO1X52hWinfoxfJWHehKamwwrq17G/V7t4JphUenGjhPBdMLmA81eK5JL9aO2l01233aqI2n+9Vze0xnqyP69eE+TVpLC8Kgf39K6+zBBv2xuk89yaQunW7VKxWN+iBmNsyqs7pRLx7u0qXbaQ12tOuP5R3qtMKjr83pCX25P6KN54Y1mLT3vfbrIWf47j1d+jritX+uVWv3d2vQH0Bd2NxDAAIQgAAEIAABCEAAAo9F4IkEzHRvnerixauXppcmYF7xVcDSfW5wTKv3mzr1/uwdy6PRK/qqbdQKbv5AadbwP7ce9/n2eXdQF852exWsWYPEIz3yLUu0u8N30+o9W6+bGa8v2aGL2usETH+/7DVm9t3d0hxv7XU7YlmvPRpV65FOWVE12Zk7Ps0RzK0KZllj3hzMz3ucPUxnNZWrEk7r0vEGJ5hJl75u0Naoe4BZXThqlrnPJfV0PEbAbNBvy+35oC+WNWjt0bgG3f2bgFnRbVdBTfemh7Q1N4TVFwatrvueD3VpbS7omoVpVR/wjkPZrKZy3Z7QlxVNskOu18ZUe5t+/VW/Nx9USR3YHdEBUwUd7tIfd3d5/VJWF75q0NZ2t+MOS+4gAAEIQAACEIAABCAAgccmEHzAfDCq1jmql6an9pBRr89eUEyoc68TvNzFmZuq//amTN3SW89e6H/uf2wvfaTxa/X66liN6puvqHfUq3y6TVv3D8ww2iv6/tta1dbWquqYGzCL9CXZmQuYM/c387jc/RQGaumubn7rhtdxxU60WnNVTWit+WHc3Szvfs4hspmkzp5r08aKiNaWN+q3ZV4wMwHTrgKa5lL6cq8bzJzm8yqVXlhT3uv+rph1fG1MZzXW2qa1BUNkve8PfG3mVSxNm75lsXatKRgCnNf38WF9WdumEnOBIesY3T54bRhGubmgTpdzbRRpv9j6/iPlMQQgAAEIQAACEIAABCCwOAKBB8z5qpeme3MHzCt2Zc89jvRN1TcOPkbAdBt4pKwZltpcXaSqmtVwS41af0wr60x69Po2rs4jdvBzW9LYlfyA6a+YmuNqP6LOIlfHNW1+/6O/Umaqo99r2Hlp/Idqtd7OarStRrGJ3N7yHsweMNM6e7hRmy4lNXnfLvHlQpXsCuaMgOmMZrV2YOZP5iqHXlhbcMC0GhnTgc+aVG2y8Yxg6mvTHyit7XzLutv1/NE+X/XRX4kd1eflTfo8nnIqtf6g7LVhBcZ6P0CnjW5JJmAez5/DOljfpD9FnIsT5dHmCQQgAAEIQAACEIAABCDwOASCDZgLqF6aTnohzu6yVw205yTa8xPNskdKdNTkgqGZL1ndPmpfTGc6rd5v9+bmZHptOBjSg/r+B3tordXSre91JFpYHXQqie7I2ulxxU65FcxHGo/WqP76pL2/R2ndjHyVC5j6uVd1pzqVeODu76YuzFK5Nce799ubcnfzaOyKqiOD3kWBfu5V7bf1qj/bm1vHaTV3N3vAdMKWGxqzt1VW7lUt/WHTNDZ4LqIXc/MSpzVYF9GapQbM4S698qkzJ3LOgDmsD8qadGDYXIRoWpMdbfq9O3z2/rA+MCGyz4xJntbUULdKcpVYEyLbdMkdIjveo41lMyuY1jDY8nZdcoc1j3brT2VtunBf0v1+bTXbjDqNZIb0QbkzfDZHmQcQgAAEIAABCEAAAhCAwFIIBBowF1K9NJ2dPWBKejCu7m+rVFVbq9pTVaq9NOyFLmvZV9p77CtV1Xbq5nV37ubM4bNmDt/wpVpVnTK/zVmjmm86NZr7sUYPmXVRof9VY/9+57lOdba4AdOsk1Xih4v2skhMiQGvgmmWps1vfzrbVv2venVPFF771d6POd4rfYO6aI7pbI2qals17KZNaxUzf3NvLkjbW+X/d/aAKU3F2rX27878zN3t+uDw7AFT0xOqPtio334a0dpPI9rY2q0PFh0wG7Sm1Lv9enebvuy/Z3d4zoA5rUHze55lDfp1WaNKIt2++ZnmukxDKjtgz+0sqR5SdW4u6bQGIy36rXWMjVp7sF1bi8zBNMHUat9Z77d/b9LnPU6/JE31RPWKw+nFv0e0qdlf7cznzTMIQAACEIAABCAAAQhAYPEEAg2Yi9/9HFs8yOaGrc6x1vyLzO9zulXGWdd+pKw7RnbWdczwz/yAaa+6wG2tlWdb1wRMd07mXB2YY9l0VpN3vDA1x5r2okzad2Ggedf+ZVbIej+vYsKi/2JFVgey9zS5kN/+nJPFQ03dyfguGPTLHBp7gQAEIAABCEAAAhCAwLNAYPkGzGVAf/xanWpbejU8Ma7xW71qra1V57+KVykft7vZiVHdjNbpq8vecN7HbWtFb5e9rc93m58Y6VNn/LbOmp8R+axdnWZ4K/8gAAEIQAACEIAABCAAgRVBgIA5z2m6OzGo3h9iil27qdGf/RfqmWfDBS7OTo5qNJn25mMucLtVuVo2pc6mLpVVd+jzhn4N3nEnXa7Ko+WgIAABCEAAAhCAAAQgsOoIEDBX3SnlgCAAAQhAAAIQgAAEIAABCDwdAgTMp8OdvUIAAhCAAAQgAAEIQAACEFh1BAiYq+6UckAQgAAEIAABCEAAAhCAAASeDgEC5tPhzl4hAAEIQAACEIAABCAAAQisOgIEzFV3SjkgCEAAAhCAAAQgAAEIQAACT4cAAfPpcGevEIAABCAAAQhAAAIQgAAEVh0BAuaqO6UcEAQgAAEIQAACEIAABCAAgadDgID5dLizVwhAAAIQgAAEIAABCEAAAquOAAFz1Z1SDggCEIAABCAAAQhAAAIQgMDTIUDAfDrc2SsEIAABCEAAAhCAAAQgAIFVR4CAuepOKQcEAQhAAAIQgAAEIAABCEDg6RAgYD4d7uwVAhCAAAQgAAEIQAACEIDAqiNAwFx1p5QDggAEIAABCEAAAhCAAAQg8HQIEDCfDnf2CgEIQAACEIAABCAAAQhAYNURCC5gTmWUyWSLAspmMspMFSx6lFUmOaDYQGLmMrNq1rTnvxVvu6BVnkIAAhCAAAQgAAEIQAACEIDAUyIQXMBMRrQjtFknBwqOZOCkNod2KJL0Xk+0VWpLOKTwxi3asuVNhUMl2n4spoy3ilIXdygUCuXf3tiuk/FfMmhmdCs2oNQvuUsfAx5CAAIQgAAEIAABCEAAAhBYSQSCC5iSEo07FPpbvRI5AgnV/y2k7ee8V7LRSoXDO1X/oy9OZgZ08r2wNh/rlZvlrID5UUSpXFtS5nqVNocPK+au5Fv2ZB72qsqE4/En0zqtQgACEIAABCAAAQhAAAIQWE0EAg2YkgmUYe1stmOhFTg3nZRX1BzQybfDqowWSYj/rNd2X5grFjAlE/g268yPvlOQuaWW/6pU+V/LVflfLbrly63uWqlYvQ5/sUM7vjis+pg/skp6lFLvxZOq+GiLyg+eUcttp2+ZXtVXV2h76G3tPHRSJ9u8kOy2yz0EIAABCEAAAhCAAAQgAAEIeAQCDpiSTFAMV6h9oEU7wwVDZq1htIfV6+1/1kdFA+btM9rur2COt6j8jc2q+LZdsVhM7d9WaHN4p1pyFcesBqq3qOS9KrXEYorFWlRlKqVfDziV0ozavwhr8956xW4ndOtqlba/sdMezptNaSB2RuWhLTrcHFPsdpHkOmvvWQABCEAAAhCAAAQgAAEIQODZIxB8wJQ0ULPZmjvpBTkHbLxKoYJhr7MhtwJm+E1t+esW57ZZJZsq1Z4Lj1LvsbDervbqo6atxLntCh+K2QFy3MwLLVfLz769TMVUGd6u+n+a15whsL75oZlUQtlH7voMkXVJcA8BCEAAAhCAAAQgAAEIQGA+Ak8gYGYVOxi2Amb4i/a8C/fIBMwPI745mppxMZ+quN1lK2D+9bBTeTTVx3ZFDm1XyZYq9VpXpE2o/q9hVfUVHOLASb39tj0s18z3DH3ckjePUzJVy5Aq2kxF0q5ghreU68ylXt2acTUfAmYBXZ5CAAIQgAAEIAABCEAAAhCYlUDgATN7/bDefvuwelOmUlgw39KqKFbOcpGehCIfhpQXMGdUO+3wuv1bMx9yQFXh/KvTWkdpDcOtsobhppp3KnR05oDc3qMh7bjozsXMKBGL6MzBcm3ZGFY4F2BNawTMWc1hAQQgAAEIQAACEIAABCAAgQICwQbMqV4d9l3EJ9NWobAJm7nfwLSvKrujscgFc6yL/FSo3ZnqWHQOpnul2mMmNPorkd5R5VUtfdVMb41ZKp/WChm17w3p7Rp32C0B0+PGIwhAAAIQgAAEIAABCEAAAnMTCDBgZjXw9WblD4tNqeXj/J8fkfW7mJtVcXFAGfeCrbdbdHhTyHfxHWforL+Cmc0oNRBRxdthHXZ+p8T6yZNNVep1r7+T6VXVprAz/NUcuHNVW1+gTTTuVNi9sq2pqIadi/qY1R8lVP+Rv7ppAuaW/KvWzs2TpRCAAAQgAAEIQAACEIAABJ5ZAoEFzGzc/Eal/wquDtNx+2qyVXHvp0mytyOq/FuJNU8zFAop9MZ2VV68lTdf06pgmmW+W8mmclVd9Vc/s7p1sVJbwiGVvFGiUPaf/ogAACAASURBVPhNlX97K/dbmlYPfu7VyY/MshKVmPX+VqWY76I/iTazfVhvbtmsklBYW/a3580RTTSXq8T0oXA+6TOrDAcOAQhAAAIQgAAEIAABCECgOIHAAmbx5ud5NZtRxi1jzrPq3IuzymQy+cGycIOpjDK5obqFC53tc1ePLVzOcwhAAAIQgAAEIAABCEAAAhCYj8DTDZjz9Y7lEIAABCAAAQhAAAIQgAAEILBiCBAwV8ypoqMQgAAEIAABCEAAAhCAAASWNwEC5vI+P/QOAhCAAAQgAAEIQAACEIDAiiFAwFwxp4qOQgACEIAABCAAAQhAAAIQWN4ECJjL+/zQOwhAAAIQgAAEIAABCEAAAiuGAAFzxZwqOgoBCEAAAhCAAAQgAAEIQGB5EyBgLu/zQ+8gAAEIQAACEIAABCAAAQisGAIEzBVzqugoBCAAAQhAAAIQgAAEIACB5U2AgLm8zw+9gwAEIAABCEAAAhCAAAQgsGIIEDBXzKmioxCAAAQgAAEIQAACEIAABJY3AQLm8j4/9A4CEIAABCAAAQhAAAIQgMCKIUDAXDGnio5CAAIQgAAEIAABCEAAAhBY3gQImMv7/NA7CEAAAhCAAAQgAAEIQAACK4YAAXPFnCo6CgEIQAACEIAABCAAAQhAYHkTIGAu7/ND7yAAAQhAAAIQgAAEIAABCKwYAmsGx++IGwxwAAdwAAdwAAdwAAdwAAdwAAeW6sCa1AOJGwxwAAdwAAdwAAdwAAdwAAdwAAeW6gABk4DNFww4gAM4gAM4gAM4gAM4gAM4EIgDBExECkSkpX7TwfZ8W4YDOIADOIADOIADOIADK98BAiYBk4CJAziAAziAAziAAziAAziAA4E4QMBEpEBE4tumlf9tE+eQc4gDOIADOIADOIADOLBUBwiYBEwCJg7gAA7gAA7gAA7gAA7gAA4E4gABE5ECEWmp33SwPd+W4QAO4AAO4AAO4MDyd6D++8v6P//7/9D/9uv//bFuZlvTBud6+Z/rxz1HBEwCJm9wHMABHMABHMABHMABHFiQAy/+t8cPl24oNSHzccML2y3/YErA5I8Jb3AcwAEcwAEcwAEcwAEcWJADbkhc6j1BcfkHxcc9RwRM/pgs6I/J4wrGdqv3jwfnlnOLAziAAziAA8+eA0sNlu72uLN63SFgEjAJmDiAAziAAziAAziAAziwIAfcgLiY+/9rw59mzNckYBIwFyQcoqxeUTi3nFscwAEcwAEcwAEcwIHFBEuz7v/zzv9r5Yj/PF6dFzJxafW6RAWTb6v48gAHcAAHcAAHcAAHcAAHFuTAYgKmGy5NmCRgrt5AWfhlAQGTPyYL+mNSKA7Pn50/EpxrzjUO4AAO4AAO4IDrwEID5lzh0rThtsf96nNrGQXMhxqbSGvo39PBCXf/js5/E1VVf3bJbY7Fe1T6zZCGgg6k97Ma+vG2aq4ngz32oPtJe0t2iD+gq+8PKOeUc4oDOIADT8+B5GRCnR0Dik8u/HNecjKla9d6dS05xzZTWQ0O9qptMKXk1GzHl9XIyICabyQ0kpltHSmVnnt/ycmMRgpv6TnaWwafxwoDZrH5lfOFSwLm8j7HS/27towC5pC2ljbo5fpUcB/kx3r0H6UNeuH06JLbjJxo0JqyVlVNBCfEWLxLr5U1aE2pcyuLaOPFiSX3dalSsH1w5xiWsMQBHMABHMCBoB1IqHHPFm1YH9I66xbWO4diGpkzfBVuE1L4wzPqLAhzyRtntK3EbTekdSXlqh7MD6Mz1llfoo++SxR8fsvqWl25wrk+FtlfJqZdvuX2sYS07mhvQVtB81tae/6AaYa9Gr9NoHRfX0i4JGAu7Rws978pqztgmj80kxmN3Q/gJJpK4+TD4N7w92/r/fIG/e5IXPG7Uup+RpfPNum50hYdSgTQ3zn/yNL+cn9j0j8cxQEcwAEcwIHiDsTrtmvd+h063p+xPpeNxKr01vqQtl2YvUgx0lSet01qMqY9b4W04WBMSfczUyamPa+G9Of97Ro0lcuphGp3hrXurSq1uVVKZ523DjrrPJDi9Tu1Yf12VQ97/U1eq9Kf12/WnstOn5Lt2vOXgv2NRbRt/RYd7ymoYhaE3uXmgRskzf3bm+wL+Jg+mmC50HBJwPRcWW7nN4j+BBsw76Z0vj6qjQcievkfHdoXu5MLZEPRLr1/Iq6o+yZ+IEWbO/T+N8POsFO3gplUtKVDr3/RpNePd6lm1At1bhuXx4a060iLtY9D8XtKWUNh2/TaZy3aaIax5gJlUlUnOrQrmnb6Ma2hgT7tOt6ql611+xSd9J3gRfV/WkM9PSo90qKXvmjV+/XDdlB0js89tqgZWmvWOdCmXVc9HqmxuDZ+FlHpNd/+b3fr5dIGbe30vebjZZ3wiVEdOm2OtUmvn+jWeR+f1IOHike78/pks3ioy9916P3T/Yr72xuI6/0TUdWMzrE///o8zvkcxJuPNvAOB3AAB3AABxbrwICOvBXSn6sG8v6f3HY0rHVb6vM/5/g+t7QdDWndhxENzvHaSEuF1q0v1zn/aDUrBIa0q92pYmZS6uyI6VqRdSpzn+kyavysIEw+kJIjvWruuOVVWq9Vad36nTqXXCyDp7u+P2AWhkzX58IL+hRuQ8B8uufQPU9P6j7AgJnSoX0Neu6zFpV+169Dp1v0QmmDXr9oh7t4fZPWlLbrvO+Nff54g9bs6Xb+GNgB83flEf2hsk3vnzAhygxLbdK+YXtept1Go17Y06JNvuWv7YvoteMdev9Ik35lhsSeuu380XFDq/3t0Vis3erTy8e7VdXSrU1fNGhNeYciViBdTP+nFa031cYGvXykw+5ruTn2qNOWZB3bJ416abfpa4c27jHDYBu1tTN/mEXeib3Wod+VRrRrYBbphrv1H2UN+tUew8dp08/nQpOeK4votdNx1XwX1WumT5VxK8CPXW7TcwVtR041as0n7vHPsk/f+crrK6/n/Y8NNviDAziAAziAA7+AA07g2xPL31eyvXLOsJbsOKwN63eoesTZLt2ryr+E9NYJL6haIfS9SEFITah6S0gbjnnrFZ5nq4L5aoUac6GzV5Xr7VA60t+i41UndaSqXo0j+Z8B7T4f1rmeiLfOsF2VLdzHcnpeLCz6K5kLCZcEzHx/l9P5DaIvwQXMRI9eK23QpituxdFU+PoVcSpsCw2Yaw70eBfSMcNIP2nQc8eHrA/zdhu+IaTOPl86O+Z82J/W+WMzQ6s7r9OaR7m7y6uiTiZ1Ppq0K56L6f+/+7WprEF/yO1XSjnbv/adHaitgFnWqhq3Qnp/VKUmMDvHMuPkTQ5pa3mDXjg+pLGi4S2rmqMNWmP671Zo749p34GI/lBrAvUdHTrQoDVH+73gMzaqmuvO0Iz7Q9pq+vxN0llus/1dLoyvbtFn8C7KGAZwwgEcwAEcwIFl7YATML1qoXO+rGrgDtWOzX7+4t9V6K31Yf3PdzcrvP5NvXuofc6Kps0hpdoPi82LTFnzQN95M6x1f6lQrf+Ckk4ft+3cqfC75dpTdVK7PizRuvWbVXnNC5mDF3ZonenPR5U6Mss6y/FcFAuY5jUTMhcaLgmYs3u6HM/5YvsUXMB8YFcATcVx4+luVV1PaszMLXQ+yC80YLph0N3OCoVfdFuhcGYb+RVKs02xqqjbplvBfKmyTbu+G9blhPcmTy2m/7F2uxrY7x1f6kFaVZUNWnOkzwqI+f0w66W0z1QxiwXMydt6/4sGPbev2wuPMwLQsBUQXz43+0WA4qaCaVVVo9rXclvRvDmj04qcbLQDqmn7+jzV0hn79x8rj10/uccFHMABHMABHPgFHXjMgGnmRL7zaljvHoyosSOmxrpK6/l7F7yL8xQbRms+vxUPmFkN9sTU3FRvh8e/VKjRrY46fVy3zT9kN6vm/UWG8eZdpXaWdZbZZ7LZAuZiX+d98wu+b35hhwIMmFLKmsNo5k826jfm6qhlEW3tvGeFzJnhsHgYdIfUutJFz0bsYZzWJOrCYbaLC5ipB84cTDMn8hP7yq2/OtDjhbqF9r+zXWtKm7Tvtl8Mp3q6r8caWrHggHk3qX3W0OJ2nXernUUlmHmsLiPv3p6D+f4/mvTS3+3je+mEryLqC5XW8Fh/NbfoPv3Hx2OPMyxggQM4gAM4gANPxYFki95zhp/692/Pn6xQY9HPUs4wV/8FfR5I9oV/vG06jxUJgA+cOZ++obT+/dqP7XVyw2idq8Nuq/fCq1nPHRLbnBcq8z2y1/H6NHNf+es/jeWLDZKzrf80+s4+fxl/Ag6YWe+Krfcn7IqdM8fSDphtOu8O73xwT1WHZg5n/d1Jd/6kAZC2h3068whnhtSZoSs/2BUun9bYv90hvNLY1Tb9zn9Rnbvz9d+ZQ+pcjGfTZV8F1BnO6w45ze+HOZYiFUzDaJ+ZBzpfuPS2f+5ov28IbVbxWL9q+tyLGD3UWO53RKcVPWcCuT8Ij2nXZ+anYOLW0GO3ssub7Zd5s8EZzjiAAziAAziwVAcSqt5WeAEdp/KXVzH078eeE/nOmfzAl4odtq4s6w6rtedp5l8NNjVcr3fXh7Wnw/7Ml+xvseZTtuXmW5r9JFT7XkjrDsackXt2HwsvRBQ/s0XrXq1Sp/WlflbXvjupI3W93kV/TDHFXCE3t47/GJbP4//jd7/P/STJbOFxvtdf/G//IzfKkffE8jm3QZ2L4ALmQJf+UNqg/zg3ZofMfw3r/d0Neu5Qnz2n0rqATYNeOBrX5dGkzp9usaucBRf5ea6sSe9HUxqaSClS3zrPhYIKA2TxqqgdpJw5jOXtqvnXQ6XuZxW/0KLnShtVGpdS8/Q/P9w6w2HL21T14z2NTaYUsX5iJKJd/fYFieYNmG64LI1o68V+1VzxbpFh+49Y/LtWvfRFu2qcny0Zumj3d1NLUkMTaUVb2nx8ktpleJthtiZk3k3r/MnIjN/ujH5jXmvg51Co2PKHHQdwAAdwAAdWpAOmWrlh/WZ9fGFAg5MpdV6w51Z+3OJeICelc59t0baj7m9jZtV8MKx1r+7U8WsJjUxmNDJi/2zIur+cdAKf+ZA/oCPmp0S2V6l5xKwTU+V2M8fSt85Euz5+NaTwx2fUZtaZTKj52HZtWB/2rjRrqqN5fcwo3lGld1/Nv6jQQtYJ6gN/kO3Uf39ZJiDOFyJnW262NW0E2SfaWl4hNbiA+WBa0YuteskMjS11hp/uiSqSG6pwT5HT9lVezfI/nBhSVZGryG5tGdJWczEcp42XT9/OVezyQ54BuZiAaX4T87ZKrau5Ou2XRbTxojunce7+z9j35Kh2VTZacx6tvpY1qTTm/mErDLqmrwUVTKcK6h6n/96tLF4+3WhfRTd3Vdl7uvyNE8yL8EkNm58+8diZ+bCl1+whyrk3nrtf/8WU+B8sf+RwAAdwAAdwAAdWjANZdZ7aofD6kNaZ26tv6qO6W97vWT64peMmKPqvCDuV0Ln9271trIvrnFFn7nOq8wF9pF273g3b7a4PKbytSs151UopORjRrm3moj3O/ku2a1dTQXX0QVbXLph5ns4660u0rSq/Wpla0DrLKzjkPk+uGFfg9zTOWYAB0z2BDzU2kdZQbqim+/pi7qc1lkrnXSQoUDj/zmhooiB45d4oi+y/aSvlGyqba2cxx7vIde9nrQqm/yJKeXwm0798n36J42YffPjBARzAARzAARxwHZjKWtXIpPt8gffJyYySc8yDNJ+pzDoj6Xk+n2VMBXP+z4DW/ubp20LWyfusN097rDvPuYPfE/078gQCJieUNzUO4AAO4AAO4AAO4AAO4AAOPIsOEDD5BuOJfoPxLL6pOGb+Z4IDOIADOIADOIADOPCsOkDAJGASMHEAB3AAB3AAB3AAB3AAB3AgEAcImIgUiEjP6jc0HDffTuIADuAADuAADuAADuCA5wABk4BJwMQBHMABHMABHMABHMABHMCBQBwgYCJSICLxrY33rQ0sYIEDOIADOIADOIADOPCsOkDAJGASMHEAB3AAB3AAB3AAB3AAB3AgEAcImIgUiEjP6jc0HDffTuIADuAADuAADuAADuCA5wABk4BJwMQBHMABHMABHMABHMABHMCBQBwgYCJSICLxrY33rQ0sYIEDOIADOIADOIADOPCsOhBYwBztb9Wx07UFt1Z1/TxTrv5orS79c+brq/UkBHG88YuteumzSP7tUFzxpQbksbg2uu30RfVSWauqJqSU//W8fdxWqb8fB9pU+t2oxtx1Zt3uSZ1v05+oIu7+uecLAxzAARzAARzAARzAARx4ag4EFjBTU1klMr7b5KDq/nFF16dmBosbbXsVGZn5+moNmEEcb7y+SWuO9mloIu3dUtmli3O7Wy/v6baD6v2Mon0pu03/63lv0CFtLW3Srh6nH6NjOnS0US/Xz7fdkzrfpj/tOp/Xxye1L9pdre9Rjgu3cQAHcAAHcAAHcCAYB4ILmAUf8G911+nrrnTRAGQFruG7ut4dU/PVbkWT+UEpMTGstqsxNUdv6oavApoY6VXbiLeu/7l53PWvrG6YNvud/WbS6nL2cfWf3naF8iSSg0X3Z9Yzyy5djelS77gS6YTa3LbN8d6/qxu93Wq+GlPbQFqJAgbufuyA+UhDfWbdbl0duWtzmZpU9OpN9ftD+L+G1XxjUuMFbVkB8/hQUZ6pB9Ma6olr14kOvX+6W+dHH+bWizZ36fzwqA6d7tCuqMNltN9e95s+ReO+gPkgqaoTcUXNvucJmPtu+wQc7tbL+3rskOpsd7mnR6WmP9/0K37XXdfXvnV8Bc/vZ3T5u6jePxHVvtgdDUW7VDXgbntP0SvdTpt9ik5MO8dYEDDvpnS+3rTRoV0tSQ3dd7c35yujaEuXtay0ftjXL986Bdzdc8g9jHAAB3AAB3AAB3AAB3BgfgeeTMCcGtWFo8Wrl+akmMB16Eynro6ldWukV3UnanQpaXd2fKRTx850qutfd+1lpy7o6k/2stHeelX3OuHsgeR/bh4fPXVRkb5RXf+XCZPjaj5Vp8hwWomfx9VW/5UuDD/KBS9XjtEbF7Tf2d/o2E3VnfheXWl7f4n+73PLbo10q76uTofaEk4baV2tr1Vdd0K3fp5UV3udjrUlZgRD93i/rmtV88CkRn9KqO1irWp6Tdh7pOst+f260Zb/3O3nXAEzfqFJv/qiXVV9KcWvd+v18hZVJexjOH+8US/sadWuln5FhrNKXe/QC+Wt2hVNKt4X16Y9jfqdW8F84AtrCw6YDxX9pkkvnB61uZjtPmnU66f6dfnHMVWdiOhXh/o0ZAU3X/sznqdVVdmgP5zo0+XRpM6fatXLXzRqa6c5jqwiJxr1wqEunf8xreiVdv2hvEMRKzz62rw/oX37Inr9m2FFR+19v3R8yBm+e0/nj0e89r9p1Uv7uhX1B1DC5Yz3h+sf9/b7CQ5wwAEcwAEcwAEcwIG5HXgiAfNWV52+7i5evTQnxATMC7e8jt3qcYNjWm21dWozcwCdD/vjw1e0v2XUCm7+QGmW+59bj3t8+/x5UDWnu70K1qxB4pHGfctutLrDd9NqO12vq5NeXxJ9F/WJEzD9/bL7OrPv7jGY4z0Wm8wdU2pqVPX/6NQNc4wjnbnjM6/PFsytgFnWmDcHs/Sa07f7WY3lqoTTOn+swQlm0vnjDdp01a32ZVVzxCxzn0upax2PETAb9Jtyez7oC2UNeulIXFF3/yZgftFtV0HN8d0f0qbcEFZfGLTOr+95f5deygVdc1xpHTrgHUfqblZjufM0oX1fNMmuonptjF1u06+O9nvzQR8ktWt3RLtMFXSgS3/Y3eX160FWNUcbtOny7JVt9/xx770HYAELHMABHMABHMABHMCBuRwIPmDeHVX9HNVL0xl7yKh3YrygmFDkCyd4OQEzNXlT1XU3NVoQKE073nb5j+0DfqT+aL32H61R9XdX1DbsVT7zgNw1w2ivqK6uVsdO1ergUTdgFunLSGcuYPr37bZXeFz+1/2BOvXgrq7WueF1XM1VrdZcVRNaj14d94Koy+CBNFcFMzWZVNU3bXr9i4heKm/Ub8q8YGYCpl0FNLxT2rfHDWYO/7xKpRfWFjxE9n5W8ZY2vVQwRNa7+JCvTX+FtDBgdrZrTcEQ4Ly+jw1r3+k2vWYuMGQdo3scXvuGUW4uqMMu10aR9out754z7r33JyxggQM4gAM4gAM4gAM4sFAHAg+Y81UvTccKg5gX1kyou2JX9txw9dNNVdcPPkbAdCV4pIQZlvpddZGqalZdF2tU359WwpkH6fVtXJF/2MEvB/PWlfyA6a+YmuNqPaJIkavjmjbr+v2VMlMd/V5dGbuP/VerVT+U1fWWGjWPuf3Ov589YKZVdahRG5uTGvq3XZnMhaoHdgVzRsAc9rVt5k/mKodeWFtwwLTO05h2fdakQ6bveYHV7MfXZt7jgmWxdj13pM9XffRXYkdVWt6k0uspp1LrD8pe+1ZgPDfhC+hOGzEpZQLmsfw5rNFzTfqPC87FiVzfuPfx83kCF7jgAA7gAA7gAA7gAA4swIFgA+YCqpdzB0x7TqI9P9F8uH2kG201uWBo5kseah21L6ZzP622ur25OZleSHU+FP80qLqr9tBas8/xge+1p72wOuhUEp05nqn742o+4VYwH6m/vUbVsUl7f1NpXW34KhcwUxO9+vpEp264Q0N/uqmaWSq3JmB+UndTt5wTMn7rig41DHoXBZro1bG6elWf7s2tkwu1zjazB0wnbLmh8e5tvV/uVS39YdO0Gf0mohdy8xKnFT0b0ZqlBsyBLr38iTMncs6AOaytZU3aNWAuQjStoStt+p07fPbfw9pqQmRPxlo21t+t13KVWBMi23TeHSI71qPXy2ZWMK1hsOXtOu8Oax7u1n+Utanm31Lq3/3aZLYZdoYHTw5pa7kzfHYBb5TC88FzwicO4AAO4AAO4AAO4AAOzHQg0IC5kOqlOQleldDuUF44vDuuS3VVOniqVsdOVOlY07AXuqxlX+mTo1/p4KlOXY25czeLDZFNq6upVgdPmN/mrNHR2k5ddy7e4xfBuqjQsRr79zvPdCpy0Q2Ypm9Z3bh60V7WENONG14F07Rxy/z2p7PtwWP1ujQ28yJC7vFe6BnUGXNMp2t08FSrutxQa4UbM39zby5I+/vnPp49YEpjne166e/O/Mzd7dp6aPaAmbo/oUOVjfrNJxG99ElEr7d0a+uiA2aD1pR6t1/tbtO++D37G505A+a0oub3PMsa9KuyRr12ods3P9P89uaQ3j9gz+187cSQDuXmkk4reqFFv7GOsVEvVbZrU5E5mCa0Wu076/3m700qveb064E0di2qlx1OL/w9oo0X/dXOmW8Olz33sMEBHMABHMABHMABHMCBhTkQaMAMFPrdbG7Y6pLaNb/P6VYZZ61UPVIiUzwc5u3bN0TWe32B21r7nm1dEzDdOZkLO3He/p3172c1NOGFqRnLC499Mu27MNBj7rOwzaU+v+v9vIoJi/6LFVnHc/eehhby259zsniosYmM74JBy+TYl8qO7RmyggM4gAM4gAM4gAM4sAwcWL4BcxnA6Y/W6djFXnWNjat/oFf1p2oV+ecCgugi+p4YG9XV9jrt/94bzjtvOFxE+yumrbu3Vbrb/MRInyLXb6vK/IzIZ+2KmOGtq/F4OSbOKw7gAA7gAA7gAA7gwCp0gIA5z0kdHRtU29WYmqM3dX3Cf6GeYIJP4l+juj6S9uZjztOfVR227qYU+a5L75/oUGl9v6ITvp9TeZa5cOz8zwcHcAAHcAAHcAAHcGCFOEDAXCEnalUHS84BfzBxAAdwAAdwAAdwAAdwYFU4QMBE5FUhMgE8mIo6HOGIAziAAziAAziAAziwFAcImARMAiYO4AAO4AAO4AAO4AAO4AAOBOIAARORAhFpKd9ysC3fkuEADuAADuAADuAADuDA6nCAgEnAJGDiAA7gAA7gAA7gAA7gAA7gQCAOEDARKRCR+MZpdXzjxHnkPOIADuAADuAADuAADizFAQImAZOAiQM4gAM4gAM4gAM4gAM4gAOBOEDARKRARFrKtxxsy7dkOIADOIADOIADOIADOLA6HCBgEjAJmDiAAziAAziAAziAAziAAzgQiAMETEQKRCS+cVod3zhxHjmPOIADOIADOIADOIADS3GAgEnAJGDiAA7gAA7gAA7gAA7gAA7gQCAOEDARKRCRlvItB9vyLRkO4AAO4AAO4AAO4AAOrA4HCJgETAImDuAADuAADuAADuAADuAADgTiAAETkQIRiW+cVsc3TpxHziMO4AAO4AAO4AAO4MBSHHhiATM5mVFyaqEnJ6uRyeyCg87i2l5oH1hvKSKxLf7gAA7gAA7gAA7gAA7gAA48mYA5FtG29SH9uWpgYaHRWr9KbQuqJqZU+2FIldee7slLjg2oeTCzsONb0HE93ePhjwH8cQAHcAAHcAAHcAAHcAAHlurAEwmY8brt2vDeDr376gJD4woMmIMXdmjd0V4CJuEZB3AAB3AAB3AAB3AAB3AABxwHnkDATKh629s6csPch7Wno8jQ18mEGusq9dGWcu2q69XIiKl45ofRkf4WHT9UoW0fVmhPXbviaffbBLeCmdW1ppP6+MMd+vhQvZqT7nLvfrCjXns+26Ftnx1WdUdqxkm39rG/XO98XKnjTbc0UvDGGBluV7VZbvWhN7c83nJSe3a+rXXbKnSkql5tk94+l5r42R6WOIADOIADOIADOIADOIADK9WB4APmjZP681sn1flAMlW+DQdjSvqDW7pXlX8J691jLWruiKmxrkLbdu7UO76AOdJSoQ2vbldlU686b/Sq9uBmbdjZokGrHTtgbttZrm2+Nt56dadqR1wRs+qs2qLw9iqd64ipuaNFldvDeuvYQK4vg03lCv+lQtWXzfJ2VX/h34eUunFSb1l91A6jHAAABbVJREFUGFB8pNda/tbRXmv7kcGYzh3aonWfnlFzx4AGM+5+uV+pbwT6jbs4gAM4gAM4gAM4gAM4sHQHAg+YnVVv688nnLmXEy36aH2FGn0VPmv47P780Bk/s13rfAEzlU7o2oiv8pmJadf6nTpnVSntgLmhSBu5MGsNuS3XuQkfoHRMu17druph81qvKl81VVbf8gf5FVdrCOwh3xDYTEaDE16fGCLrZ8dj/hjhAA7gAA7gAA7gAA7gAA5IAQdME9y2OCHOCJbRuY9D+qjJvRhOVs37Q/q4xX3uSDhcn1fBtE5MJqP4jZiam+p15FC53lq/Q7VjZn07YM5ow1c5TbZXal2u4umKnlHjZ86+zf5erbKqrH4JOk/4wrGpYK4v0bajETXeSGikoEpJwHS5cu93iMf4gAM4gAM4gAM4gAM48Cw7EGjATHYc1ob1Ia0rvH0YyQ1vPbczpF3tXiXQgl8wBzP+XbnCr76p9/af1JG6FjXfiGhPQcDcEysQ19fG4Hc7tc5ffXSG6LYdCmnbhZRSPVXa8F5Ecf/Q3QdSvL7gwj3JAZ2rO6yPP9ys8PoSffRdIjePk4BZwL+A5bP8puLYcQMHcAAHcAAHcAAHcOBZdSDAgJlV88Gw3v2vAY1MZrxbskUfr3eHpkqdx8LeEFonlFgVx9wQ2QEdeatg+Ko1RDa/gvnOGS/smZOXV7X0VTO9E5tQ9ZawKnuk1GS7Pi4Yupt6YFdX3/tu5sWArDb6z+gd3zYETP5oeG7BAhY4gAM4gAM4gAM4gAM4YBwILmBaoc0Lkp5gTvCscwKhdfEc3wV5JntV+W5YG3IB0wyBDWvXZXcYbVbXzuzUhoIK5oaScu+iPqaNv4R9Q2/t+ZTv1XshNF6/Uxv+Yl98yA6TYZmL9rhXjh2JVemtVyvU6MzbbDtaYi13L1A00nFYf/b97IoVML9oz23vHS9iwQIHcAAHcAAHcAAHcAAHcODZdCCwgDnSVK51uaGw+TCtobPOlWWNaPGWSr3zakjhN9+UCYrnYvk/U5K8cUbbSkLa8OZm/c9XzdDUmUNkKy/bwdRqw8yVPOVdIdaSeaJXRz4s0bpXSxQ2+9pWpea8i/7cUu2eLdqwvkRha1/lqh70Dd1ND+i42b5ks955M6x1JTt0/IZ/ub3/desLqq0MFc0NI+aPSv77AB7wwAEcwAEcwAEcwAEcWO0OBBYwFw8qaw2jdSuExbZPTmaUnJpbwnnXSWc0kvsNzSJtZcxwXl9wLAyIZvu5lheuz3MCJg7gAA7gAA7gAA7gAA7gwDPqwFMMmEXC3jN6EoqFa17DDxzAARzAARzAARzAARzAgZXmAAGTUMu3SziAAziAAziAAziAAziAAzgQiAMETEQKRKSV9s0K/eXbQBzAARzAARzAARzAARwI3gECJgGTgIkDOIADOIADOIADOIADOIADgThAwESkQETi25/gv/2BKUxxAAdwAAdwAAdwAAdWmgMETAImARMHcAAHcAAHcAAHcAAHcAAHAnGAgIlIgYi00r5Zob98G4gDOIADOIADOIADOIADwTtAwCRgEjBxAAdwAAdwAAdwAAdwAAdwIBAHCJiIFIhIfPsT/Lc/MIUpDuAADuAADuAADuDASnOAgEnAJGDiAA7gAA7gAA7gAA7gAA7gQCAOEDARKRCRVto3K/SXbwNxAAdwAAdwAAdwAAdwIHgHCJgETAImDuAADuAADuAADuAADuAADgTiAAETkQIRiW9/gv/2B6YwxQEcwAEcwAEcwAEcWGkOEDAJmARMHMABHMABHMABHMABHMABHAjEAQImIgUi0kr7ZoX+8m0gDuAADuAADuAADuAADgTvwJp//jQpbjDAARzAARzAARzAARzAARzAARxYqgNrxD8IQAACEIAABCAAAQhAAAIQgEAABAiYAUCkCQhAAAIQgAAEIAABCEAAAhCQ/n9QPcHn3TJjrwAAAABJRU5ErkJggg=="
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Adaboost model with hyperparameter tuning performns better and is also take less time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "293px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}