웹에서 사용자에게 보여줄 그래프를 그리기 위해 interactive한 그래프를 그려주는 plotly를 활용한다.
import pandas as pd
import plotly.graph_objects as go
from plotly.offline import plot
import database.db as dbt
import required.data as reqdata그래프를 그리기 위해 데이터베이스에 저장된 데이터를 불러온다.
sql_call = dbt.DBcls.sqlselectALL()
db_data = pd.DataFrame(sql_call)
projects_fin = reqdata.data_fin
temp_data = projects_fin[
['name', 'state_new', 'main_category', 'deadline', 'launched', 'usd_pledged_real', 'usd_goal_real', 'term',
'term_bin', 'usd_goal_real_bin', 'backers', 'currency']].copy()
temp_data.rename(columns={'main_category': 'category', 'state_new': 'state'}, inplace=True)
data = pd.concat([db_data, temp_data], sort=False)VisCls class는 웹에 띄울 그래프를 그려주는 5개의 함수를 담고 있는 그래프이다. VisCls 객체가 생성되면 사용자가 입력한 프로젝트의 카테고리 정보, 펀딩 기간 정보 (전처리 로직에 따라 binning), 목표 금액 정보 (전처리 로직에 따라 binning)가 전역 변수로 설정된다.
class VisCls():
def __init__(self, cate, term_bin, usd_goal_real_bin):
self.cate = cate
self.term_bin = term_bin
self.usd_goal_real_bin = usd_goal_real_bin카테고리와 목표 금액이 주어졌을 때 '기간별' 성공률을 확인할 수 있는 함수이다. 주어진 카테고리와 목표 금액 값에 따라 binning된 기간에 해당하는 전체 프로젝트 수를 분모로, 그 중에 성공한 프로젝트 수를 분자로 하여 성공률을 계산한다.
def g_success_term(self):
df = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
percent_list = []
for i in range(1, 9):
numerator = len(df.loc[(df['usd_goal_real_bin'] == i) & (df['state'] == 'success')])
denominator = len(df.loc[df['usd_goal_real_bin'] == i])
print("Numerator:", numerator, "Denominator:", denominator)
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': ['below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 3000', 'above 3000~<br>below 5000', 'above 5000~<br>below 10000', 'above 10000~<br>below 50000', 'above 50000~<br>below 100000', 'above 100000'],
'y': percent_list}
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days and below 92 days']
title_main = "Rate of success for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, yaxis=dict(tickformat="%"), xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Goal amount(USD)")), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str카테고리와 펀딩 기간이 주어졌을 때 목표 금액이 어떻게 분포되어있는지, 목표 금액별로 성공한 프로젝트와 실패한 프로젝트의 비율이 어떻게 되는지를 확인할 수 있는 함수이다.
def g_goal_dist(self):
df = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
x_range = ['below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 3000', 'above 3000~<br>below 5000', 'above 5000~<br>below 10000', 'above 10000~<br>below 50000', 'above 50000~<br>below 100000', 'above 100000']
fin = df.groupby(['state', 'usd_goal_real_bin']).size().to_frame().unstack().fillna(0)
fin.index.names = [None]
fin = fin[0].T.reset_index()
####################
# DATA
####################
suc = go.Bar(name='success', x=x_range, y=fin['success'], marker_color='#1100FF')
fail = go.Bar(name='failed', x=x_range, y=fin['failed'], marker_color='#FE0002')
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days']
title = "Number of projects for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
####################
# GRAPH
####################
fig = go.Figure(data=[suc, fail])
fig.update_layout(barmode='stack', title={'text': title}, title_x=0.5, font={'size': 9}, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Goal amount(USD)")))
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str카테고리와 펀딩 기간이 주어졌을 때 후원자 1인당 투자 금액의 분포를 확인할 수 있다. 이 때 1인당 투자 금액은 USD로 환산된 값을 기준으로 한다.
def g_pledge_pp(self):
df_call = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
df = df_call.loc[df_call['backers'] != 0]
df = df.copy()
df['pledge_per_person'] = df['usd_goal_real'] / df['backers']
df['pledge_per_person_bin'] = df.apply(lambda x: '1' if x['pledge_per_person'] <= 50 else '2' if x['pledge_per_person'] <= 100 else '3' if x['pledge_per_person'] <= 500 else '4' if x['pledge_per_person'] <= 1000 else '5' if x['pledge_per_person'] <= 2500 else '6' if x['pledge_per_person'] <= 5000 else '7', axis=1)
percent_list = []
for i in range(7):
sval = str(i + 1)
numerator = len(df.loc[df['pledge_per_person_bin'] == sval])
denominator = len(df)
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': ['below 50', 'above 50~<br>below 100', 'above 100~<br>below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 2500', 'above 2500~<br>below 5000', 'above 5000'],
'y': percent_list}
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days']
title_main = "Distribution for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Average funding amount(USD) per person")), yaxis=dict(tickformat="%"), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str카테고리와 목표 금액이 주어졌을 때 기간별로 성공률이 어떻게 되는지 확인한다. 1번의 함수는 카테고리와 기간을 가지고 목표 금액별 성공률을 시각화했다면, 이 함수는 카테고리와 목표 금액을 가지고 기간별 성공률을 확인했다는 점에서 차이가 있다. 모델링 결과 가장 중요한 변수로 지정된 카테고리와 목표 금액, 기간을 충분히 활용하여 시각화를 시도하였다.
def g_success_cat_amount(self):
cate_df = data[lambda x: (x['category'] == self.cate) & (x['usd_goal_real_bin'] == self.usd_goal_real_bin)]
percent_list = []
# 각 기간별 성공률 확인
for i in range(1, 9):
numerator = len(cate_df[lambda x: (x['term_bin'] == i) & (x['state'] == 'success')])
denominator = len(cate_df[lambda x: x['term_bin'] == i])
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
# x축은 기간, y축은 성공률
res_dct = {'type': 'bar',
'x': ['10 days or less', 'above 10 days~<br> below 15 days', 'above 15 days~<br> below 21 days', 'above 21 days~<br> below 30 days', 'above 30 days~<br> below 45 days', 'above 45 days~<br> below 60 days',
'above 60 days'],
'y': percent_list}
####################
# LABELS
####################
goal_list = ['below $500', 'above $500 ~ below $1000', 'above $1000 ~ below $3000', 'above $3000 ~ below $5000', 'above $5000 ~ below $10000',
'above $10000 ~ below $50000', 'above $50000 ~ below $100000', 'above $100000']
title_main = "Rate of success for '"+ self.cate +"'<br> when funding goal is " + goal_list[self.usd_goal_real_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Funding period")),yaxis=dict(tickformat="%"),font={'size':9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str카테고리, 목표 금액, 기간이 주어졌을 때 어떤 화폐단위가 가장 성공률이 높은지를 확인할 수 있는 함수이다. 화폐 단위는 웹에서 받아올 수 있는 단위인 AUD, CAD, EUR, GBP, SGD, USD만 가지고 그래프로 나타내었다.
def g_success_by_curr(self):
cate_df = data[
lambda x: (x['category'] == self.cate) & (x['usd_goal_real_bin'] == self.usd_goal_real_bin) & (
x['term_bin'] == self.term_bin)]
curr_list = ['AUD', 'CAD', 'EUR', 'GBP', 'SGD', 'USD']
percent_list = []
for i in curr_list:
numerator = len(cate_df[lambda x: (x['currency'] == i) & (x['state'] == 'success')])
denominator = len(cate_df[lambda x: x['currency'] == i])
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': curr_list,
'y': percent_list}
####################
# LABELS
####################
goal_list = 'below $500', 'above $500 ~ below $1000', 'above $1000 ~ below $3000', 'above $3000 ~ below $5000', 'above $5000 ~ below $10000', 'above $10000 ~ below $50000', 'above $50000 ~ below $100000', 'above $100000'
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days', 'above 60 days']
title_main = "Rate of success for '" + self.cate +"<br> when funding period is " + term_list[
self.term_bin - 1] + "<br> and funding goal is " + goal_list[self.usd_goal_real_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Type of Currency")), yaxis=dict(tickformat="%"), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_strimport pandas as pd
import plotly.graph_objects as go
from plotly.offline import plot
import database.db as dbt
import required.data as reqdata
##############################
# Visual
##############################2
sql_call = dbt.DBcls.sqlselectALL()
db_data = pd.DataFrame(sql_call)
projects_fin = reqdata.data_fin
temp_data = projects_fin[
['name', 'state_new', 'main_category', 'deadline', 'launched', 'usd_pledged_real', 'usd_goal_real', 'term',
'term_bin', 'usd_goal_real_bin', 'backers', 'currency']].copy()
temp_data.rename(columns={'main_category': 'category', 'state_new': 'state'}, inplace=True)
data = pd.concat([db_data, temp_data], sort=False)
class VisCls():
def __init__(self, cate, term_bin, usd_goal_real_bin):
self.cate = cate
self.term_bin = term_bin
self.usd_goal_real_bin = usd_goal_real_bin
def g_success_term(self):
df = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
percent_list = []
for i in range(1, 9):
numerator = len(df.loc[(df['usd_goal_real_bin'] == i) & (df['state'] == 'success')])
denominator = len(df.loc[df['usd_goal_real_bin'] == i])
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': ['below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 3000', 'above 3000~<br>below 5000', 'above 5000~<br>below 10000', 'above 10000~<br>below 50000', 'above 50000~<br>below 100000', 'above 100000'],
'y': percent_list}
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days']
title_main = "Rate of success for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, yaxis=dict(tickformat="%"), xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Goal amount(USD)")), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str
def g_goal_dist(self):
df = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
x_range = ['below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 3000', 'above 3000~<br>below 5000', 'above 5000~<br>below 10000', 'above 10000~<br>below 50000', 'above 50000~<br>below 100000', 'above 100000']
fin = df.groupby(['state', 'usd_goal_real_bin']).size().to_frame().unstack().fillna(0)
fin.index.names = [None]
fin = fin[0].T.reset_index()
####################
# DATA
####################
suc = go.Bar(name='success', x=x_range, y=fin['success'], marker_color='#1100FF')
fail = go.Bar(name='failed', x=x_range, y=fin['failed'], marker_color='#FE0002')
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days']
title = "Number of projects for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
####################
# GRAPH
####################
fig = go.Figure(data=[suc, fail])
fig.update_layout(barmode='stack', title={'text': title}, title_x=0.5, font={'size': 9}, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Goal amount(USD)")))
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str
def g_pledge_pp(self):
df_call = data.loc[(data['category'] == self.cate) & (data['term_bin'] == self.term_bin)]
df = df_call.loc[df_call['backers'] != 0]
df = df.copy()
df['pledge_per_person'] = df['usd_goal_real'] / df['backers']
df['pledge_per_person_bin'] = df.apply(lambda x: '1' if x['pledge_per_person'] <= 50 else '2' if x['pledge_per_person'] <= 100 else '3' if x['pledge_per_person'] <= 500 else '4' if x['pledge_per_person'] <= 1000 else '5' if x['pledge_per_person'] <= 2500 else '6' if x['pledge_per_person'] <= 5000 else '7', axis=1)
percent_list = []
for i in range(7):
sval = str(i + 1)
numerator = len(df.loc[df['pledge_per_person_bin'] == sval])
denominator = len(df)
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': ['below 50', 'above 50~<br>below 100', 'above 100~<br>below 500', 'above 500~<br>below 1000', 'above 1000~<br>below 2500', 'above 2500~<br>below 5000', 'above 5000'],
'y': percent_list}
####################
# LABELS
####################
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days',
'above 60 days']
title_main = "Distribution for '"+ self.cate +"'<br> when funding period is " + term_list[self.term_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Average funding amount(USD) per person")), yaxis=dict(tickformat="%"), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str
def g_success_cat_amount(self):
########################################################################
# 카테고리와 목표 금액이 주어졌을 때 기간별로 성공률이 어떻게 되는지 확인한다.
########################################################################
cate_df = data[lambda x: (x['category'] == self.cate) & (x['usd_goal_real_bin'] == self.usd_goal_real_bin)]
percent_list = []
# 각 기간별 성공률 확인
for i in range(1, 9):
numerator = len(cate_df[lambda x: (x['term_bin'] == i) & (x['state'] == 'success')])
denominator = len(cate_df[lambda x: x['term_bin'] == i])
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
# x축은 기간, y축은 성공률
res_dct = {'type': 'bar',
'x': ['10 days or less', 'above 10 days~<br> below 15 days', 'above 15 days~<br> below 21 days', 'above 21 days~<br> below 30 days', 'above 30 days~<br> below 45 days', 'above 45 days~<br> below 60 days',
'above 60 days'],
'y': percent_list}
####################
# LABELS
####################
goal_list = ['below $500', 'above $500 ~ below $1000', 'above $1000 ~ below $3000', 'above $3000 ~ below $5000', 'above $5000 ~ below $10000',
'above $10000 ~ below $50000', 'above $50000 ~ below $100000', 'above $100000']
title_main = "Rate of success for '"+ self.cate +"'<br> when funding goal is " + goal_list[self.usd_goal_real_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Funding period")),yaxis=dict(tickformat="%"),font={'size':9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str
def g_success_by_curr(self):
# 카테고리, 목표 금액, 기간이 주어졌을 때 어떤 화폐단위가 가장 성공률이 높은가
cate_df = data[
lambda x: (x['category'] == self.cate) & (x['usd_goal_real_bin'] == self.usd_goal_real_bin) & (
x['term_bin'] == self.term_bin)]
curr_list = ['AUD', 'CAD', 'EUR', 'GBP', 'SGD', 'USD']
percent_list = []
for i in curr_list:
numerator = len(cate_df[lambda x: (x['currency'] == i) & (x['state'] == 'success')])
denominator = len(cate_df[lambda x: x['currency'] == i])
if denominator == 0:
percent = 0
else:
percent = round(((numerator / denominator)), 2)
percent_list.append(percent)
####################
# DATA
####################
res_dct = {'type': 'bar',
'x': curr_list,
'y': percent_list}
####################
# LABELS
####################
goal_list = 'below $500', 'above $500 ~ below $1000', 'above $1000 ~ below $3000', 'above $3000 ~ below $5000', 'above $5000 ~ below $10000', 'above $10000 ~ below $50000', 'above $50000 ~ below $100000', 'above $100000'
term_list = ['10 days or less', 'above 10 days and below 15 days', 'above 15 days and below 21 days', 'above 21 days and below 30 days', 'above 30 days and below 45 days', 'above 45 days and below 60 days', 'above 60 days']
title_main = "Rate of success for '" + self.cate +"<br> when funding period is " + term_list[
self.term_bin - 1] + "<br> and funding goal is " + goal_list[self.usd_goal_real_bin - 1]
title = {'text': title_main}
####################
# GRAPH
####################
fig = go.Figure({"data": res_dct, "layout": {"title": title}})
fig.update_traces(marker_color='#78BD40')
fig.update_layout(title_x=0.5, xaxis=go.layout.XAxis(title=go.layout.xaxis.Title(text="Type of Currency")), yaxis=dict(tickformat="%"), font={'size': 9})
html_str = plot(fig, output_type='div', include_plotlyjs=False)
return html_str