admariner / eda_user_churn

Exploratory Data Analysis for Churn prediction. Python vs BigQuery SQL.

Medium Article:

Exploratory Data Analysis with BigQuery SQL? Easy! -- by Mike Shakhomirov

---

Customer Retention

Aims

The aim of this project is to:

• Analyse Churn dataset
• Analyze how the different features affect retention or customer churn
• Compare Python and SQL techniques

Data

The dataset used for this project was obtained from Kaggle. It contains data about customers who are withdrawing their account from a bank.

Files

This repo contains the following files:

• A jupyter notebook named churn_analysis.ipynb containing the exploratory data analysis, feature engineering, search for the best model, evaluations of the best models found, as well as the analysis of the feature importance.

• SQL queries to do the same in BigQuery: churn_analysis.sql

• A CSV file named Churn.csv which contains the data from the bank customers