Cohort Retention Analysis

Setup

import pandas as pd import matplotlib.pyplot as plt import warnings import seaborn as sns from operator import attrgetter import matplotlib.colors as mcolors import numpy as np

df = pd.read_excel('Online Retail.xlsx', dtype={'CustomerID': str, 'InvoiceID': str}, engine='openpyxl', parse_dates=['InvoiceDate']) df.head()

Exploratory analysis

df.info()

# drop rows without customer ID df2 = df.dropna(subset=['CustomerID'], inplace=False)

df2.head()

df2.describe().transpose()

# how many orders were placed by each customer? n_orders = df2.groupby('CustomerID')['InvoiceNo'].nunique() mult_orders_percent = np.sum(n_orders > 1) / df2['CustomerID'].nunique() print(f'{100 * mult_orders_percent:.2f}% of customers ordered more than once.')

# look at distribution of the number of orders per customer ax = sns.distplot(n_orders, kde=False, hist=True) ax.set_title('Distribution of the number of orders per customer') ax.set(xlabel='Number of orders', ylabel='Number of customers')

Cohort retention analysis

df2 = df2[['InvoiceNo', 'CustomerID', 'InvoiceDate']].drop_duplicates()

# create cohort and order month variables df2['order_month'] = df2['InvoiceDate'].dt.to_period('M') df2['cohort'] = df2.groupby('CustomerID')['order_month'].transform('min')

# count the number of unique customers in each cohort and order month df_cohort = df2.groupby(['cohort', 'order_month']).agg(n_customers=('CustomerID', 'nunique')).reset_index(drop=False) # find number of months between the cohort month and the month of the purchase df_cohort['period_num'] = (df_cohort['order_month'] - df_cohort['cohort']).apply(attrgetter('n'))

# pivot df_cohort on cohort cohort_pivot = df_cohort.pivot(index='cohort', columns='period_num', values='n_customers') cohort_pivot

# divide number of customers in each subsequent period by cohort size cohort_size = cohort_pivot.iloc[:, 0] retention_matrix = cohort_pivot.divide(cohort_size, axis=0)

# plot retention matrix as heatmap with sns.axes_style("white"): fig, ax = plt.subplots(1, 2, figsize=(12, 8), sharey=True, gridspec_kw={'width_ratios': [1, 11]}) # retention matrix sns.heatmap(retention_matrix, mask=retention_matrix.isnull(), annot=True, fmt='.0%', cmap='RdYlGn', ax=ax[1]) ax[1].set_title('Monthly Cohorts: User Retention', fontsize=16) ax[1].set(xlabel='# of periods', ylabel='') # cohort size cohort_size_df = pd.DataFrame(cohort_size).rename(columns={0: 'cohort_size'}) white_cmap = mcolors.ListedColormap(['white']) sns.heatmap(cohort_size_df, annot=True, cbar=False, fmt='g', cmap=white_cmap, ax=ax[0]) fig.tight_layout()