feature_engineering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import seaborn as sns\n",
    "sns.set(color_codes = True)\n",
    "import numpy as np\n",
    "from sklearn import preprocessing\n",
    "import imblearn\n",
    "from imblearn.combine import SMOTETomek"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#member_v3 = pd.read_csv('members_v3.csv')\n",
    "#train = pd.read_csv('train.csv')\n",
    "transaction = pd.read_csv('transactions.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3 = pd.read_csv('members_v3.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>city</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rb9UwLQTrxzBVwCB6+bCcSQWZ9JiNLC9dXtM1oEsZA8=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>20110911</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>+tJonkh+O1CA796Fm5X60UMOtB6POHAwPjbTRVl/EuU=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7</td>\n",
       "      <td>20110914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cV358ssn7a0f7jZOwGNWS07wCKVqxyiImJUX6xcIwKw=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>20110915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9bzDeJP6sQodK73K5CBlJ6fgIQzPeLnRl0p5B77XP+g=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>20110915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WFLY3s7z4EZsieHCt63XrsdtfTEmJ+2PnnKLH5GY4Tk=</td>\n",
       "      <td>6</td>\n",
       "      <td>32</td>\n",
       "      <td>female</td>\n",
       "      <td>9</td>\n",
       "      <td>20110915</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  city  bd  gender  \\\n",
       "0  Rb9UwLQTrxzBVwCB6+bCcSQWZ9JiNLC9dXtM1oEsZA8=     1   0     NaN   \n",
       "1  +tJonkh+O1CA796Fm5X60UMOtB6POHAwPjbTRVl/EuU=     1   0     NaN   \n",
       "2  cV358ssn7a0f7jZOwGNWS07wCKVqxyiImJUX6xcIwKw=     1   0     NaN   \n",
       "3  9bzDeJP6sQodK73K5CBlJ6fgIQzPeLnRl0p5B77XP+g=     1   0     NaN   \n",
       "4  WFLY3s7z4EZsieHCt63XrsdtfTEmJ+2PnnKLH5GY4Tk=     6  32  female   \n",
       "\n",
       "   registered_via  registration_init_time  \n",
       "0              11                20110911  \n",
       "1               7                20110914  \n",
       "2              11                20110915  \n",
       "3              11                20110915  \n",
       "4               9                20110915  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                            0\n",
       "city                            0\n",
       "bd                              0\n",
       "gender                    4429505\n",
       "registered_via                  0\n",
       "registration_init_time          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.isnull().sum(axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "missing_values = pd.DataFrame()\n",
    "missing_values = member_v3.isnull().sum(axis = 0).reset_index()\n",
    "missing_values.columns = ['variable', 'missing_count']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x1d12b3bcc50>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhIAAAFlCAYAAAC+8gFbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2v8HIV9CwV9D8j4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d00ee1b940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.barplot( x = 'variable', y = 'missing_count', data = missing_values)\n",
    "#plt.xticks(rotation = 90)\n",
    "plt.ylabel('missing_count')\n",
    "plt.title('missing values analysis column wise')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def change_datatype(df):\n",
    "    int_cols = list(df.select_dtypes(include=['int']).columns)\n",
    "    for col in int_cols:\n",
    "        if ((np.max(df[col]) <= 127) and(np.min(df[col] >= -128))):\n",
    "            df[col] = df[col].astype(np.int8)\n",
    "        elif ((np.max(df[col]) <= 32767) and(np.min(df[col] >= -32768))):\n",
    "            df[col] = df[col].astype(np.int16)\n",
    "        elif ((np.max(df[col]) <= 2147483647) and(np.min(df[col] >= -2147483648))):\n",
    "            df[col] = df[col].astype(np.int32)\n",
    "        else:\n",
    "            df[col] = df[col].astype(np.int64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "change_datatype(member_v3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "309.8819580078125  MB\n"
     ]
    }
   ],
   "source": [
    "mem = member_v3.memory_usage(index=True).sum()\n",
    "print(mem/ 1024**2,\" MB\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                      object\n",
       "city                       int64\n",
       "bd                         int64\n",
       "gender                    object\n",
       "registered_via             int64\n",
       "registration_init_time     int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3['registration_init_time'] = pd.to_datetime(member_v3['registration_init_time'],\n",
    "                                                   format='%Y%m%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_month(x):\n",
    "    txt = str(x)\n",
    "    y = txt.split('-')\n",
    "    if len(y) != 3:\n",
    "        return 'NaN'\n",
    "    return int(y[1])\n",
    "\n",
    "def get_year(x):\n",
    "    txt = str(x)\n",
    "    y = txt.split('-')\n",
    "    if len(y) != 3:\n",
    "        return 'NaN'\n",
    "    return int(y[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3['month'] = member_v3['registration_init_time'].map(lambda x: get_month(x))\n",
    "member_v3['year'] = member_v3['registration_init_time'].map(lambda x: get_year(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>city</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "      <th>month</th>\n",
       "      <th>year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rb9UwLQTrxzBVwCB6+bCcSQWZ9JiNLC9dXtM1oEsZA8=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>2011-09-11</td>\n",
       "      <td>9</td>\n",
       "      <td>2011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>+tJonkh+O1CA796Fm5X60UMOtB6POHAwPjbTRVl/EuU=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7</td>\n",
       "      <td>2011-09-14</td>\n",
       "      <td>9</td>\n",
       "      <td>2011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cV358ssn7a0f7jZOwGNWS07wCKVqxyiImJUX6xcIwKw=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>2011-09-15</td>\n",
       "      <td>9</td>\n",
       "      <td>2011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>9bzDeJP6sQodK73K5CBlJ6fgIQzPeLnRl0p5B77XP+g=</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11</td>\n",
       "      <td>2011-09-15</td>\n",
       "      <td>9</td>\n",
       "      <td>2011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>WFLY3s7z4EZsieHCt63XrsdtfTEmJ+2PnnKLH5GY4Tk=</td>\n",
       "      <td>6</td>\n",
       "      <td>32</td>\n",
       "      <td>female</td>\n",
       "      <td>9</td>\n",
       "      <td>2011-09-15</td>\n",
       "      <td>9</td>\n",
       "      <td>2011</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  city  bd  gender  \\\n",
       "0  Rb9UwLQTrxzBVwCB6+bCcSQWZ9JiNLC9dXtM1oEsZA8=     1   0     NaN   \n",
       "1  +tJonkh+O1CA796Fm5X60UMOtB6POHAwPjbTRVl/EuU=     1   0     NaN   \n",
       "2  cV358ssn7a0f7jZOwGNWS07wCKVqxyiImJUX6xcIwKw=     1   0     NaN   \n",
       "3  9bzDeJP6sQodK73K5CBlJ6fgIQzPeLnRl0p5B77XP+g=     1   0     NaN   \n",
       "4  WFLY3s7z4EZsieHCt63XrsdtfTEmJ+2PnnKLH5GY4Tk=     6  32  female   \n",
       "\n",
       "   registered_via registration_init_time  month  year  \n",
       "0              11             2011-09-11      9  2011  \n",
       "1               7             2011-09-14      9  2011  \n",
       "2              11             2011-09-15      9  2011  \n",
       "3              11             2011-09-15      9  2011  \n",
       "4               9             2011-09-15      9  2011  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZkAAAEFCAYAAAAxAZr2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x168180c43c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(6,4))\n",
    "sns.countplot(x = 'gender',data = member_v3)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3['gender'] = member_v3['gender'].fillna('missing')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "missing    4429505\n",
       "male       1195355\n",
       "female     1144613\n",
       "Name: gender, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.gender.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2223464\n",
      "6769473\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "32.845451928089531"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print (member_v3['bd'][(member_v3['bd'] > 0) & (member_v3['bd'] < 100)].count())\n",
    "print (len(member_v3['bd']))\n",
    "(member_v3['bd'][(member_v3['bd'] > 0) & (member_v3['bd'] < 100)].count())*100/len(member_v3['bd'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "274\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.004047582433669504"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(member_v3['bd'][member_v3['bd'] < 0].count())\n",
    "float(member_v3['bd'][member_v3['bd'] < 0].count()*100)/len(member_v3['bd'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5377\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.07943011221109826"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(member_v3['bd'][member_v3['bd'] > 100].count())\n",
    "float(member_v3['bd'][member_v3['bd'] > 100].count()*100)/len(member_v3['bd'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4540215\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "67.06895795285689"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(member_v3['bd'][member_v3['bd'] == 0].count())\n",
    "float(member_v3['bd'][member_v3['bd'] == 0].count()*100)/len(member_v3['bd'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3.loc[member_v3['bd'] < 0, 'bd'] = np.nan\n",
    "member_v3.loc[member_v3['bd'] > 100, 'bd'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "102\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([   0.,    1.,    2.,    3.,    4.,    5.,    6.,    7.,    8.,\n",
       "          9.,   10.,   11.,   12.,   13.,   14.,   15.,   16.,   17.,\n",
       "         18.,   19.,   20.,   21.,   22.,   23.,   24.,   25.,   26.,\n",
       "         27.,   28.,   29.,   30.,   31.,   32.,   33.,   34.,   35.,\n",
       "         36.,   37.,   38.,   39.,   40.,   41.,   42.,   43.,   44.,\n",
       "         45.,   46.,   47.,   48.,   49.,   50.,   51.,   52.,   53.,\n",
       "         54.,   55.,   56.,   57.,   58.,   59.,   60.,   61.,   62.,\n",
       "         63.,   64.,   65.,   66.,   67.,   68.,   69.,   70.,   71.,\n",
       "         72.,   73.,   74.,   75.,   76.,   77.,   78.,   79.,   80.,\n",
       "         81.,   82.,   83.,   84.,   85.,   86.,   87.,   88.,   89.,\n",
       "         90.,   91.,   92.,   93.,   94.,   95.,   96.,   97.,   98.,\n",
       "         99.,  100.,   nan])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print (len(np.sort(member_v3['bd'].unique())))\n",
    "np.sort(member_v3['bd'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgkAAAGACAYAAAAnNfF1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1d011ad4c18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=[8,6])\n",
    "ax.set_title(\"Distribution of age\")\n",
    "ax.set_ylabel(\"Frequency\")\n",
    "ax.set_xlabel(\"Age\")\n",
    "member_v3['bd'].hist()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2016    2246761\n",
       "2015    1620525\n",
       "2014     975776\n",
       "2013     524722\n",
       "2017     481684\n",
       "2012     283190\n",
       "2011     179051\n",
       "2010     115075\n",
       "2007      89830\n",
       "2008      67690\n",
       "2009      63633\n",
       "2006      53953\n",
       "2005      41349\n",
       "2004      26234\n",
       "Name: year, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "member_v3.year.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "14\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014,\n",
       "       2015, 2016, 2017], dtype=int64)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print (len(np.sort(member_v3['year'].unique())))\n",
    "np.sort(member_v3['year'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3.month.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "member_v3.registered_via.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "transaction.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "transaction['transaction_date'] = pd.to_datetime(transaction['transaction_date'],\n",
    "                                                   format='%Y%m%d')\n",
    "transaction['membership_expire_date'] = pd.to_datetime(transaction['membership_expire_date'],\n",
    "                                                   format='%Y%m%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(21547746, 9)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transaction.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "transaction.is_auto_renew.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mem = transaction.memory_usage(index=True).sum()\n",
    "print(mem/ 1024**2,\" MB\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "notes"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                      0\n",
       "payment_method_id         0\n",
       "payment_plan_days         0\n",
       "plan_list_price           0\n",
       "actual_amount_paid        0\n",
       "is_auto_renew             0\n",
       "transaction_date          0\n",
       "membership_expire_date    0\n",
       "is_cancel                 0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transaction.isnull().sum(axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.merge(left = train,right = transaction ,how = 'left',on=['msno'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15883148, 10)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "del transaction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161107</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>20170107</td>\n",
       "      <td>20170206</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161221</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160715</td>\n",
       "      <td>20160821</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160831</td>\n",
       "      <td>20161021</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  is_churn  payment_method_id  \\\n",
       "0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "\n",
       "   payment_plan_days  plan_list_price  actual_amount_paid  is_auto_renew  \\\n",
       "0                  7                0                   0              0   \n",
       "1                 30              149                 149              0   \n",
       "2                 30              149                 149              1   \n",
       "3                 30              149                 149              1   \n",
       "4                 30              149                 149              1   \n",
       "\n",
       "   transaction_date  membership_expire_date  is_cancel  \n",
       "0          20161031                20161107          0  \n",
       "1          20170107                20170206          0  \n",
       "2          20161031                20161221          0  \n",
       "3          20160715                20160821          0  \n",
       "4          20160831                20161021          0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    15197577\n",
       "1      685571\n",
       "Name: is_churn, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.is_churn.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.merge(left = train,right = member_v3 ,how = 'left',on=['msno'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15883148, 15)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "      <th>city</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161107</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>20170107</td>\n",
       "      <td>20170206</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161221</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160715</td>\n",
       "      <td>20160821</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160831</td>\n",
       "      <td>20161021</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  is_churn  payment_method_id  \\\n",
       "0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "\n",
       "   payment_plan_days  plan_list_price  actual_amount_paid  is_auto_renew  \\\n",
       "0                  7                0                   0              0   \n",
       "1                 30              149                 149              0   \n",
       "2                 30              149                 149              1   \n",
       "3                 30              149                 149              1   \n",
       "4                 30              149                 149              1   \n",
       "\n",
       "   transaction_date  membership_expire_date  is_cancel  city    bd  gender  \\\n",
       "0          20161031                20161107          0  18.0  36.0  female   \n",
       "1          20170107                20170206          0  18.0  36.0  female   \n",
       "2          20161031                20161221          0  10.0  38.0    male   \n",
       "3          20160715                20160821          0  10.0  38.0    male   \n",
       "4          20160831                20161021          0  10.0  38.0    male   \n",
       "\n",
       "   registered_via  registration_init_time  \n",
       "0             9.0              20050406.0  \n",
       "1             9.0              20050406.0  \n",
       "2             9.0              20050407.0  \n",
       "3             9.0              20050407.0  \n",
       "4             9.0              20050407.0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    15197577\n",
       "1      685571\n",
       "Name: is_churn, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.is_churn.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                       object\n",
       "is_churn                    int64\n",
       "payment_method_id           int64\n",
       "payment_plan_days           int64\n",
       "plan_list_price             int64\n",
       "actual_amount_paid          int64\n",
       "is_auto_renew               int64\n",
       "transaction_date            int64\n",
       "membership_expire_date      int64\n",
       "is_cancel                   int64\n",
       "city                      float64\n",
       "bd                        float64\n",
       "gender                     object\n",
       "registered_via            float64\n",
       "registration_init_time    float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "male      3595401\n",
       "female    3247448\n",
       "Name: gender, dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.gender.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                            0\n",
       "is_churn                        0\n",
       "payment_method_id               0\n",
       "payment_plan_days               0\n",
       "plan_list_price                 0\n",
       "actual_amount_paid              0\n",
       "is_auto_renew                   0\n",
       "transaction_date                0\n",
       "membership_expire_date          0\n",
       "is_cancel                       0\n",
       "city                      1600752\n",
       "bd                        1600752\n",
       "gender                    9040299\n",
       "registered_via            1600752\n",
       "registration_init_time    1600752\n",
       "dtype: int64"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.isnull().sum(axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['city'].fillna(9999.0, inplace=True)\n",
    "train['bd'].fillna(4444444.0, inplace = True)\n",
    "train['gender'].fillna('missing', inplace = True)\n",
    "train['registered_via'].fillna(9999.0, inplace = True)\n",
    "train['registration_init_time'].fillna(99999999.0, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train =  train.loc[(train.city != 9999.0)]\n",
    "train =  train.loc[(train.bd != 4444444.0)]\n",
    "train =  train.loc[(train.gender != 'missing')]\n",
    "train =  train.loc[(train.registered_via != 9999.0)]\n",
    "train =  train.loc[(train.registration_init_time != 99999999.0)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6842849, 15)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    6491838\n",
       "1     351011\n",
       "Name: is_churn, dtype: int64"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.is_churn.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                       object\n",
       "is_churn                    int64\n",
       "payment_method_id           int64\n",
       "payment_plan_days           int64\n",
       "plan_list_price             int64\n",
       "actual_amount_paid          int64\n",
       "is_auto_renew               int64\n",
       "transaction_date            int64\n",
       "membership_expire_date      int64\n",
       "is_cancel                   int64\n",
       "city                      float64\n",
       "bd                        float64\n",
       "gender                     object\n",
       "registered_via            float64\n",
       "registration_init_time    float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.to_csv('train_new.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "391692"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train['msno'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.loc[ train['gender'] == 'male', 'gender'] = 1\n",
    "train.loc[ train['gender'] == 'female', 'gender'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = train.copy()\n",
    "y = train['is_churn']\n",
    "X = X.drop('is_churn', axis = 1)\n",
    "X = X.drop('msno', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'pd' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-1-b443a69414ec>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mtrain\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mread_csv\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'train_new'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'pd' is not defined"
     ]
    }
   ],
   "source": [
    "train = pd.read_csv('train_new')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\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>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "      <th>city</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161107</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>20170107</td>\n",
       "      <td>20170206</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161221</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160715</td>\n",
       "      <td>20160821</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160831</td>\n",
       "      <td>20161021</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0                                          msno  is_churn  \\\n",
       "0           0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1   \n",
       "1           1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1   \n",
       "2           2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1   \n",
       "3           3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1   \n",
       "4           4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1   \n",
       "\n",
       "   payment_method_id  payment_plan_days  plan_list_price  actual_amount_paid  \\\n",
       "0                 38                  7                0                   0   \n",
       "1                 38                 30              149                 149   \n",
       "2                 39                 30              149                 149   \n",
       "3                 39                 30              149                 149   \n",
       "4                 39                 30              149                 149   \n",
       "\n",
       "   is_auto_renew  transaction_date  membership_expire_date  is_cancel  city  \\\n",
       "0              0          20161031                20161107          0  18.0   \n",
       "1              0          20170107                20170206          0  18.0   \n",
       "2              1          20161031                20161221          0  10.0   \n",
       "3              1          20160715                20160821          0  10.0   \n",
       "4              1          20160831                20161021          0  10.0   \n",
       "\n",
       "     bd  gender  registered_via  registration_init_time  \n",
       "0  36.0  female             9.0              20050406.0  \n",
       "1  36.0  female             9.0              20050406.0  \n",
       "2  38.0    male             9.0              20050407.0  \n",
       "3  38.0    male             9.0              20050407.0  \n",
       "4  38.0    male             9.0              20050407.0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Unnamed: 0', 'msno', 'is_churn', 'payment_method_id',\n",
       "       'payment_plan_days', 'plan_list_price', 'actual_amount_paid',\n",
       "       'is_auto_renew', 'transaction_date', 'membership_expire_date',\n",
       "       'is_cancel', 'city', 'bd', 'gender', 'registered_via',\n",
       "       'registration_init_time'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = train.drop('Unnamed: 0', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "      <th>city</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161107</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>20170107</td>\n",
       "      <td>20170206</td>\n",
       "      <td>0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050406.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20161031</td>\n",
       "      <td>20161221</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160715</td>\n",
       "      <td>20160821</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>20160831</td>\n",
       "      <td>20161021</td>\n",
       "      <td>0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>20050407.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  is_churn  payment_method_id  \\\n",
       "0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "\n",
       "   payment_plan_days  plan_list_price  actual_amount_paid  is_auto_renew  \\\n",
       "0                  7                0                   0              0   \n",
       "1                 30              149                 149              0   \n",
       "2                 30              149                 149              1   \n",
       "3                 30              149                 149              1   \n",
       "4                 30              149                 149              1   \n",
       "\n",
       "   transaction_date  membership_expire_date  is_cancel  city    bd  gender  \\\n",
       "0          20161031                20161107          0  18.0  36.0  female   \n",
       "1          20170107                20170206          0  18.0  36.0  female   \n",
       "2          20161031                20161221          0  10.0  38.0    male   \n",
       "3          20160715                20160821          0  10.0  38.0    male   \n",
       "4          20160831                20161021          0  10.0  38.0    male   \n",
       "\n",
       "   registered_via  registration_init_time  \n",
       "0             9.0              20050406.0  \n",
       "1             9.0              20050406.0  \n",
       "2             9.0              20050407.0  \n",
       "3             9.0              20050407.0  \n",
       "4             9.0              20050407.0  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6842849, 15)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0572410702033612"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train.msno.unique())/train.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "391692"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train.msno.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['registration_init_time'] = pd.to_datetime(train['registration_init_time'],\n",
    "                                                   format='%Y%m%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['transaction_date'] = pd.to_datetime(train['transaction_date'],\n",
    "                                                   format='%Y%m%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['membership_expire_date'] = pd.to_datetime(train['membership_expire_date'],\n",
    "                                                   format='%Y%m%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['trans_month'] = train['transaction_date'].map(lambda x: get_month(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['trans_year'] = train['transaction_date'].map(lambda x: get_year(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['expire_month'] = train['membership_expire_date'].map(lambda x: get_month(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['expire_year'] = train['membership_expire_date'].map(lambda x: get_year(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['registration_month'] = train['registration_init_time'].map(lambda x: get_month(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['registration_year'] = train['registration_init_time'].map(lambda x: get_year(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "trans_dates = train['transaction_date'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4oAAAHTCAYAAABhk+buAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x17da3f67320>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trans_dates.plot(kind =\"line\",figsize=(15,8))\n",
    "plt.title(\"Trend over Time - Transaction Dates\")\n",
    "plt.xlabel(\"Transaction Dates\")\n",
    "plt.ylabel(\"No Of Transactions\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "      <th>...</th>\n",
       "      <th>bd</th>\n",
       "      <th>gender</th>\n",
       "      <th>registered_via</th>\n",
       "      <th>registration_init_time</th>\n",
       "      <th>trans_month</th>\n",
       "      <th>trans_year</th>\n",
       "      <th>expire_month</th>\n",
       "      <th>expire_year</th>\n",
       "      <th>registration_month</th>\n",
       "      <th>registration_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>2016-11-07</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2005-04-06</td>\n",
       "      <td>10</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>2017-01-07</td>\n",
       "      <td>2017-02-06</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>36.0</td>\n",
       "      <td>female</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2005-04-06</td>\n",
       "      <td>1</td>\n",
       "      <td>2017</td>\n",
       "      <td>2</td>\n",
       "      <td>2017</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>2016-12-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2005-04-07</td>\n",
       "      <td>10</td>\n",
       "      <td>2016</td>\n",
       "      <td>12</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-07-15</td>\n",
       "      <td>2016-08-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2005-04-07</td>\n",
       "      <td>7</td>\n",
       "      <td>2016</td>\n",
       "      <td>8</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-08-31</td>\n",
       "      <td>2016-10-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>38.0</td>\n",
       "      <td>male</td>\n",
       "      <td>9.0</td>\n",
       "      <td>2005-04-07</td>\n",
       "      <td>8</td>\n",
       "      <td>2016</td>\n",
       "      <td>10</td>\n",
       "      <td>2016</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  is_churn  payment_method_id  \\\n",
       "0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "\n",
       "   payment_plan_days  plan_list_price  actual_amount_paid  is_auto_renew  \\\n",
       "0                  7                0                   0              0   \n",
       "1                 30              149                 149              0   \n",
       "2                 30              149                 149              1   \n",
       "3                 30              149                 149              1   \n",
       "4                 30              149                 149              1   \n",
       "\n",
       "  transaction_date membership_expire_date  is_cancel        ...            bd  \\\n",
       "0       2016-10-31             2016-11-07          0        ...          36.0   \n",
       "1       2017-01-07             2017-02-06          0        ...          36.0   \n",
       "2       2016-10-31             2016-12-21          0        ...          38.0   \n",
       "3       2016-07-15             2016-08-21          0        ...          38.0   \n",
       "4       2016-08-31             2016-10-21          0        ...          38.0   \n",
       "\n",
       "   gender registered_via  registration_init_time trans_month  trans_year  \\\n",
       "0  female            9.0              2005-04-06          10        2016   \n",
       "1  female            9.0              2005-04-06           1        2017   \n",
       "2    male            9.0              2005-04-07          10        2016   \n",
       "3    male            9.0              2005-04-07           7        2016   \n",
       "4    male            9.0              2005-04-07           8        2016   \n",
       "\n",
       "   expire_month  expire_year  registration_month  registration_year  \n",
       "0            11         2016                   4               2005  \n",
       "1             2         2017                   4               2005  \n",
       "2            12         2016                   4               2005  \n",
       "3             8         2016                   4               2005  \n",
       "4            10         2016                   4               2005  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "790"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['transaction_date'].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6842849, 21)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.loc[train['bd'] < 0, 'bd'] = 1000000\n",
    "train.loc[train['bd'] > 100, 'bd'] = 1000000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train =  train.loc[(train.bd != 1000000)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['bd'] = train['bd'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6835311, 21)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[    0  -149  -150   149  -894    30  -129  -119  -799    20   180 -1788\n",
      "  -536  -500   -35    -1  -100  -300    50  -134  -480  -350     1   120\n",
      "  -447  -450  -400  -930 -1599   -50  -105 -1200   -30 -1150  -131 -1000\n",
      "  -760   699]\n"
     ]
    }
   ],
   "source": [
    "train['discount'] = train['plan_list_price'] - train['actual_amount_paid']\n",
    "print(train.discount.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1]\n"
     ]
    }
   ],
   "source": [
    "train['is_discount'] = train.discount.apply(lambda x: 1 if x > 0 else 0)\n",
    "print(train.is_discount.unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['amt_per_day'] = train['actual_amount_paid'] / train['payment_plan_days']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#--- difference in days ---\n",
    "train['membership_duration'] = train.membership_expire_date - train.transaction_date\n",
    "train['membership_duration'] = train['membership_duration'] / np.timedelta64(1, 'D')\n",
    "train['membership_duration'] = train['membership_duration'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['registration_duration'] = train.membership_expire_date - train.registration_init_time\n",
    "train['registration_duration'] = train['registration_duration'] / np.timedelta64(1, 'D')\n",
    "train['registration_duration'] = train['registration_duration'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['reg_mem_duration'] = train['registration_duration'] - train['membership_duration']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1], dtype=int64)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['autorenew_&_not_cancel'] = ((train.is_auto_renew == 1) == (train.is_cancel == 0)).astype(np.int8)\n",
    "train['autorenew_&_not_cancel'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1], dtype=int64)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['notAutorenew_&_cancel'] = ((train.is_auto_renew == 0) == (train.is_cancel == 1)).astype(np.int8)\n",
    "train['notAutorenew_&_cancel'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['msno', 'is_churn', 'payment_method_id', 'payment_plan_days',\n",
       "       'plan_list_price', 'actual_amount_paid', 'is_auto_renew',\n",
       "       'transaction_date', 'membership_expire_date', 'is_cancel', 'city', 'bd',\n",
       "       'gender', 'registered_via', 'registration_init_time', 'trans_month',\n",
       "       'trans_year', 'expire_month', 'expire_year', 'registration_month',\n",
       "       'registration_year', 'discount', 'is_discount', 'amt_per_day',\n",
       "       'membership_duration', 'autorenew_&_not_cancel',\n",
       "       'notAutorenew_&_cancel', 'registration_duration', 'reg_mem_duration'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>msno</th>\n",
       "      <th>is_churn</th>\n",
       "      <th>payment_method_id</th>\n",
       "      <th>payment_plan_days</th>\n",
       "      <th>plan_list_price</th>\n",
       "      <th>actual_amount_paid</th>\n",
       "      <th>is_auto_renew</th>\n",
       "      <th>transaction_date</th>\n",
       "      <th>membership_expire_date</th>\n",
       "      <th>is_cancel</th>\n",
       "      <th>...</th>\n",
       "      <th>registration_month</th>\n",
       "      <th>registration_year</th>\n",
       "      <th>discount</th>\n",
       "      <th>is_discount</th>\n",
       "      <th>amt_per_day</th>\n",
       "      <th>membership_duration</th>\n",
       "      <th>autorenew_&amp;_not_cancel</th>\n",
       "      <th>notAutorenew_&amp;_cancel</th>\n",
       "      <th>registration_duration</th>\n",
       "      <th>reg_mem_duration</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>2016-11-07</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4233</td>\n",
       "      <td>4226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=</td>\n",
       "      <td>1</td>\n",
       "      <td>38</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>0</td>\n",
       "      <td>2017-01-07</td>\n",
       "      <td>2017-02-06</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.966667</td>\n",
       "      <td>30</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4324</td>\n",
       "      <td>4294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-10-31</td>\n",
       "      <td>2016-12-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.966667</td>\n",
       "      <td>51</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4276</td>\n",
       "      <td>4225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-07-15</td>\n",
       "      <td>2016-08-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.966667</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4154</td>\n",
       "      <td>4117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=</td>\n",
       "      <td>1</td>\n",
       "      <td>39</td>\n",
       "      <td>30</td>\n",
       "      <td>149</td>\n",
       "      <td>149</td>\n",
       "      <td>1</td>\n",
       "      <td>2016-08-31</td>\n",
       "      <td>2016-10-21</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4</td>\n",
       "      <td>2005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.966667</td>\n",
       "      <td>51</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4215</td>\n",
       "      <td>4164</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           msno  is_churn  payment_method_id  \\\n",
       "0  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "1  waLDQMmcOu2jLDaV1ddDkgCrB/jl6sD66Xzs0Vqax1Y=         1                 38   \n",
       "2  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "3  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "4  QA7uiXy8vIbUSPOkCf9RwQ3FsT8jVq2OxDr8zqa7bRQ=         1                 39   \n",
       "\n",
       "   payment_plan_days  plan_list_price  actual_amount_paid  is_auto_renew  \\\n",
       "0                  7                0                   0              0   \n",
       "1                 30              149                 149              0   \n",
       "2                 30              149                 149              1   \n",
       "3                 30              149                 149              1   \n",
       "4                 30              149                 149              1   \n",
       "\n",
       "  transaction_date membership_expire_date  is_cancel        ...         \\\n",
       "0       2016-10-31             2016-11-07          0        ...          \n",
       "1       2017-01-07             2017-02-06          0        ...          \n",
       "2       2016-10-31             2016-12-21          0        ...          \n",
       "3       2016-07-15             2016-08-21          0        ...          \n",
       "4       2016-08-31             2016-10-21          0        ...          \n",
       "\n",
       "   registration_month  registration_year discount  is_discount amt_per_day  \\\n",
       "0                   4               2005        0            0    0.000000   \n",
       "1                   4               2005        0            0    4.966667   \n",
       "2                   4               2005        0            0    4.966667   \n",
       "3                   4               2005        0            0    4.966667   \n",
       "4                   4               2005        0            0    4.966667   \n",
       "\n",
       "   membership_duration  autorenew_&_not_cancel  notAutorenew_&_cancel  \\\n",
       "0                    7                       0                      0   \n",
       "1                   30                       0                      0   \n",
       "2                   51                       1                      1   \n",
       "3                   37                       1                      1   \n",
       "4                   51                       1                      1   \n",
       "\n",
       "   registration_duration  reg_mem_duration  \n",
       "0                   4233              4226  \n",
       "1                   4324              4294  \n",
       "2                   4276              4225  \n",
       "3                   4154              4117  \n",
       "4                   4215              4164  \n",
       "\n",
       "[5 rows x 29 columns]"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                       object\n",
       "is_churn                    int64\n",
       "payment_method_id           int64\n",
       "payment_plan_days           int64\n",
       "plan_list_price             int64\n",
       "actual_amount_paid          int64\n",
       "is_auto_renew               int64\n",
       "is_cancel                   int64\n",
       "city                      float64\n",
       "bd                          int32\n",
       "gender                     object\n",
       "registered_via            float64\n",
       "trans_month                 int64\n",
       "trans_year                  int64\n",
       "expire_month                int64\n",
       "expire_year                 int64\n",
       "registration_month          int64\n",
       "registration_year           int64\n",
       "discount                    int64\n",
       "is_discount                 int64\n",
       "amt_per_day               float64\n",
       "membership_duration         int32\n",
       "autorenew_&_not_cancel       int8\n",
       "notAutorenew_&_cancel        int8\n",
       "registration_duration       int32\n",
       "reg_mem_duration            int32\n",
       "dtype: object"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "datetime_cols = list(train.select_dtypes(include=['datetime64[ns]']).columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = train.drop(datetime_cols, axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.to_csv('train_new.csv', index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lbl = preprocessing.LabelEncoder()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lbl.fit(train['msno'])\n",
    "train['msno'] = lbl.transform(train['msno'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.loc[ train['gender'] == 'male', 'gender'] = 1\n",
    "train.loc[ train['gender'] == 'female', 'gender'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lbl.fit(train['gender'])\n",
    "train['gender'] = lbl.transform(train['gender'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['city'] = train['city'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['registered_via'] = train['registered_via'].astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train['amt_per_day'].fillna(9999999.0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train =  train.loc[(train.amt_per_day != 9999999.0)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                      0\n",
       "is_churn                  0\n",
       "payment_method_id         0\n",
       "payment_plan_days         0\n",
       "plan_list_price           0\n",
       "actual_amount_paid        0\n",
       "is_auto_renew             0\n",
       "is_cancel                 0\n",
       "city                      0\n",
       "bd                        0\n",
       "gender                    0\n",
       "registered_via            0\n",
       "trans_month               0\n",
       "trans_year                0\n",
       "expire_month              0\n",
       "expire_year               0\n",
       "registration_month        0\n",
       "registration_year         0\n",
       "discount                  0\n",
       "is_discount               0\n",
       "amt_per_day               0\n",
       "membership_duration       0\n",
       "autorenew_&_not_cancel    0\n",
       "notAutorenew_&_cancel     0\n",
       "registration_duration     0\n",
       "reg_mem_duration          0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.isnull().sum(axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "msno                        int64\n",
       "is_churn                    int64\n",
       "payment_method_id           int64\n",
       "payment_plan_days           int64\n",
       "plan_list_price             int64\n",
       "actual_amount_paid          int64\n",
       "is_auto_renew               int64\n",
       "is_cancel                   int64\n",
       "city                        int64\n",
       "bd                          int64\n",
       "gender                      int64\n",
       "registered_via              int64\n",
       "trans_month                 int64\n",
       "trans_year                  int64\n",
       "expire_month                int64\n",
       "expire_year                 int64\n",
       "registration_month          int64\n",
       "registration_year           int64\n",
       "discount                    int64\n",
       "is_discount                 int64\n",
       "amt_per_day               float64\n",
       "membership_duration         int64\n",
       "autorenew_&_not_cancel      int64\n",
       "notAutorenew_&_cancel       int64\n",
       "registration_duration       int64\n",
       "reg_mem_duration            int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "391244"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.msno.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('train_new.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6834422, 26)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = train.copy()\n",
    "y = train['is_churn']\n",
    "X = X.drop('is_churn', axis = 1)\n",
    "X = X.drop('amt_per_day', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "smt = SMOTETomek(ratio='auto')\n",
    "X_smt, y_smt = smt.fit_sample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_2d_space(X, y, label='Classes'):   \n",
    "    colors = ['#1F77B4', '#FF7F0E']\n",
    "    markers = ['o', 's']\n",
    "    for l, c, m in zip(np.unique(y), colors, markers):\n",
    "        plt.scatter(\n",
    "            X[y==l, 0],\n",
    "            X[y==l, 1],\n",
    "            c=c, label=l, marker=m\n",
    "        )\n",
    "    plt.title(label)\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12883720, 24)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_smt.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'target'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-8-88553e70fb07>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mX_smt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtarget\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0my_smt\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'target'"
     ]
    }
   ],
   "source": [
    "X_smt.target = y_smt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'head'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-9-e9b6cfb69096>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mX_smt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'head'"
     ]
    }
   ],
   "source": [
    "X_smt.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  3.70670000e+05,   3.80000000e+01,   7.00000000e+00, ...,\n",
       "          0.00000000e+00,   4.23300000e+03,   4.22600000e+03],\n",
       "       [  3.70670000e+05,   3.80000000e+01,   3.00000000e+01, ...,\n",
       "          0.00000000e+00,   4.32400000e+03,   4.29400000e+03],\n",
       "       [  1.72325000e+05,   3.90000000e+01,   3.00000000e+01, ...,\n",
       "          1.00000000e+00,   4.27600000e+03,   4.22500000e+03],\n",
       "       ..., \n",
       "       [  1.48294000e+05,   3.80000000e+01,   3.00000000e+01, ...,\n",
       "          0.00000000e+00,   3.59553788e+02,   3.29553788e+02],\n",
       "       [  1.70742000e+05,   3.80000000e+01,   3.00000000e+01, ...,\n",
       "          0.00000000e+00,   1.96378222e+03,   1.93378222e+03],\n",
       "       [  1.39861098e+04,   3.80000000e+01,   3.00000000e+01, ...,\n",
       "          0.00000000e+00,   8.76637477e+02,   8.46637477e+02]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_smt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'numpy.ndarray' object has no attribute 'unique'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-15-a3b730d89823>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0my_smt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0munique\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'unique'"
     ]
    }
   ],
   "source": [
    "y_smt.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeEAAAFXCAYAAACV2fZmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10dbe46e748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_2d_space(X_smt, y_smt, 'SMOTE + Tomek links')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from imblearn.under_sampling import TomekLinks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tl = TomekLinks(return_indices=True, ratio='majority')\n",
    "X_tl, y_tl, id_tl = tl.fit_sample(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6685959, 24)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_tl.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6685959"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y_tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6685959"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y_tl == 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_tl[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "c = 0\n",
    "n = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i in range(len(y_tl)):\n",
    "    if y_tl[i] == 1:\n",
    "        c = c+1\n",
    "    elif y_tl[i] == 0:\n",
    "        n = n+1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(350439, 6335520)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "c,n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}