About

This canvas is a comprehensive end-to-end churn prediction and user retention analytics platform that builds predictive models, identifies behavioral segments, and delivers data-driven retention strategies. The workflow loads user event data, engineers behavioral features, trains three different ML models (HistGradientBoosting, LogisticRegression, GradientBoosting) on clean, leakage-free feature sets, and combines the results with unsupervised K-means clustering to segment users into 4-6 behavioral cohorts, each with personalized retention strategies and actionable insights based on churn drivers like link-clicking and AI agent usage.