This project is an end-to-end

This project is an end-to-end machine learning analysis designed to predict churn (customer attrition) and calculate Customer Lifetime Value (CLV) on an e-commerce platform. Using a dataset containing 5,630 customers, this project integrates a data pipeline ranging from cleaning and RFM segmentation to predictive modeling to optimize customer retention strategies.

Key Findings :

High Model Performance: The optimized LightGBM algorithm achieved a ROC-AUC score of 0.9994.

Cost Efficiency: The model successfully achieved 0 False Negatives (no churners missed) at the optimized cost threshold.

Financial Impact: Retention strategy simulations show a potential campaign ROI of 1,007.5% with total net profits reaching 170,514.