Customer Churn Prediction & Retention Analytics – Machine Learning Project

Developed an end-to-end customer churn prediction and retention analytics system to help a telecom client identify high-risk customers before they leave and improve retention strategies.

Project highlights:

Goal: Predict which customers are likely to churn based on their usage behavior, demographics, and payment patterns.

Data Cleaning & EDA: Cleaned and explored large customer datasets to identify churn drivers — such as call duration, data usage, and overdue payments.

Feature Engineering: Created behavioral metrics (avg. recharge frequency, complaint ratio, payment delay score) that significantly improved model interpretability.

Modeling: Trained and tuned Logistic Regression, Random Forest, and XGBoost models, achieving 85% recall for churned customers.

Analytics Dashboard: Built an interactive Tableau dashboard summarizing churn trends, retention opportunities, and at-risk segments.

Business Impact: Helped the company design proactive retention offers, reducing monthly churn rate by 12% within the pilot period.