About

This canvas implements a comprehensive production-ready churn prediction system using temporal backtesting across 24 months of customer data. It generates synthetic customer data, engineers 36+ advanced features (RFM segmentation, CLV metrics, engagement decay, risk indicators), trains a Gradient Boosting classifier achieving 80%+ AUC across all time periods, and delivers business-critical outputs including intervention strategy recommendations with ROI analysis, risk-segmented customer scoring, actionable dashboards, and a validated API endpoint for real-time predictions.