About

End-to-end analysis of 409,287 Zerve user events to predict engagement and reduce churn. Includes EDA, funnel analysis, K-Means user segmentation (4 personas), Gradient Boosting churn prediction (AUC: 0.945), and a deployed Streamlit dashboard. Key finding: running code delivers a 5.0x engagement lift, and the AI Agent is Zerve's most popular entry point.