Data-Driven.py

# -*- coding: utf-8 -*-
"""Case study Data-Driven

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1sBdA1czDkbksqjeTGy4-xbIaH8vAB5j3
"""

! pip install -q lifetimes

import pandas as pd
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report
from sklearn.preprocessing import LabelEncoder
from lifetimes.datasets import load_cdnow_summary, load_cdnow_summary_data_with_monetary_value
from lifetimes import BetaGeoFitter
from lifetimes import GammaGammaFitter
from lifetimes.plotting import plot_frequency_recency_matrix, plot_probability_alive_matrix, plot_period_transactions

df = pd.read_csv('/content/churn.csv')
df.head()

print(df.info())
print (df.describe())

"""#Customer Churn"""

df = df.drop(columns=['RowNumber', 'CustomerId', 'Surname'])

le = LabelEncoder()

le.fit(df['Geography'])
df['Geography'] = le.transform(df['Geography'])

le.fit(df['Gender'])
df['Gender'] = le.transform(df['Gender'])

x = df.drop(columns=['Exited'])
y = df['Exited']
x.head()
y.head()

X_train, X_test, y_train, y_test = train_test_split(x, y)
 rf_model = RandomForestClassifier()
 rf_model.fit(X_train, y_train)

y_pred = rf_model.predict(X_test)
 print(classification_report(y_test, y_pred))

result = X_test.copy()
result['Exited'] = y_test
result['Predicted'] = y_pred
result.head()

"""#Customer Lifetime Value (CLV)"""

df = load_cdnow_summary(index_col=[0])
df.head()

bgf = BetaGeoFitter()
bgf.fit(df['frequency'], df['recency'], df['T'])

plot_frequency_recency_matrix(bgf)

plot_probability_alive_matrix(bgf)

t = 1
df['predicted_purchases'] = bgf.conditional_expected_number_of_purchases_up_to_time(t, df['frequency'], df['recency'], df['T'])
df.head()

plot_period_transactions(bgf)

dfm = load_cdnow_summary_data_with_monetary_value()
dfm = dfm[dfm['monetary_value'] > 0]
dfm.corr()

print(dfm.columns)

ggf = GammaGammaFitter()
ggf.fit(dfm['frequency'], dfm['monetary_value'])

ggf.conditional_expected_average_profit(dfm['frequency'], dfm['monetary_value'])