Churn Predictor

Built an end-to-end ML pipeline to predict customer churn for a telecom company:

Data Exploration - Analyzed 7,043 customer records; found ~26% churn rate

ML Model - Trained an XGBoost classifier with class weighting to handle imbalanced data

Model Performance - Achieved 78% accuracy, 70% recall for churners (improved from 52%)

Feature Analysis - Identified top churn predictors (Contract type, Tenure, Charges, etc.)

Business Impact - Quantified model value: $106,900 savings, 260 customers saved, $10,100 wasted on retention offers

Skills Demonstrated

Python: pandas, seaborn, matplotlib, numpy

Machine Learning: XGBoost, train-test splitting, classification metrics

Data Handling: Missing values, one-hot encoding, feature engineering

Business Analytics: Cost-benefit analysis, ROI calculation