About

This is a comprehensive Zerve user analytics and success prediction platform. The canvas ingests 409K events from 6,165 Zerve users, engineers 22+ behavioral features, trains ML models (GBM, Logistic Regression, K-Means) to identify power users vs dormant observers, detects the AHA moment (agent_worker_created, p=0.0049), quantifies the create→run gap, calculates retention curves, and generates actionable per-user dashboards with personalized recommendations—all culminating in a dark-themed Streamlit explorer app that surfaces success scores, segment membership, event journeys, and growth interventions.