About

This canvas is a comprehensive **Zerve user behavioral analytics and success scoring engine** that ingests raw hackathon event logs, engineers multi-dimensional features around user engagement depth, workflow sophistication, AI adoption, and retention, then trains an interpretable ML model (HistGradientBoosting) to predict user success and segment users into tiers (Champions, Power Users, Builders, Explorers). The workflow flows left-to-right from data loading and schema exploration → feature engineering → multi-pillar success score calculation (depth, sophistication, collaboration/AI, retention) → quantile-based tier assignment → visualizations of score distributions and pillar breakdowns → predictive ML modeling with SHAP-style feature importance and tier behavioral profiling → advanced visualizations and early-stage cohort analysis (Week 1 onboarding patterns, KMeans clustering, retention trajectories).