About

This project analyzes user interaction event data from the Zerve platform to
identify behavioral patterns that predict long-term successful usage.

The workflow includes:
- Cleaning and aggregating raw event-level data into user-level features
- Exploratory Data Analysis (EDA) using statistical summaries and visualizations
- Careful definition of user success based on sustained engagement
- Identification and mitigation of data leakage
- Interpretable machine learning using Logistic Regression
- Model evaluation using accuracy, precision, recall, F1-score, ROC-AUC,
confusion matrix, and precision-recall curves

Key findings show that sustained and consistent engagement is the strongest
driver of user success, providing actionable insights for improving onboarding
and retention strategies.

All analysis is fully reproducible within Zerve and follows competition
guidelines.