Matplotlib 102 - Basic Introduction to Multiplot, Subplot and Gridspec

matplotlib-102-dataprep.py

# ---------- IMPORT PACKAGES ----------

# Dataframe manipulation
import pandas as pd

# Matrices operation
import numpy as np

# Data viz with matplotlib
import matplotlib
import matplotlib.pyplot as plt
from matplotlib import style

# Check packages' version
print('pandas     ', pd.__version__)
print('numpy      ', np.__version__)
print('matplotlib ', matplotlib.__version__)
print('seaborn    ', sns.__version__)

# ---------- LOAD DATA SET ----------

# Load the data set into Python
df = pd.read_csv(filepath_or_buffer = 'data/WA_Fn-UseC_-Telco-Customer-Churn.csv', sep = ';')

# Print top 5 rows
df.head(n = 5)

# Metadata
df.info()

# Change column types
df = df.astype({'SeniorCitizen': object})

# ---------- DATA PREPARATION ----------

# Number of missing values in a data frame
df.isnull().sum()

# Summary statistics
df.describe()

# Check the unique values of categorical columns
for col in df.select_dtypes('object').columns:
    print(df[str(col)].value_counts(), '\n')

# Data aggregation
df_group_1 = df.groupby('PaymentMethod')[['customerID']].count().reset_index()

# ---------- AGGREGATE DATA FOR SAMPLES ----------

# 1 Number of customer by payment method
df_group_1 = df.groupby('PaymentMethod')[['customerID']].count().reset_index()
df_group_1.sort_values(by = 'customerID', ascending = False, inplace = True)
df_group_1['CummulativePerc'] = (df_group_1['customerID'].cumsum() / df_group_1['customerID'].sum())
df_group_1['CummulativeSum'] = df_group_1['customerID'].cumsum()

# 2 Number of customer by gender
df_group_2 = df.groupby('gender')[['customerID']].count().reset_index()

# 3 Number of customer by senior citizen status
df_group_3 = df.groupby('Contract')[['customerID']].count().reset_index()

# 4 Number of customer by paperless billing status
df_group_4 = df.groupby('PaperlessBilling')[['customerID']].count().reset_index()

# 5 Number of customer by churn status
df_group_5 = df.groupby('Churn')[['customerID']].count().reset_index()