About

This project analyzes user behavior to identify patterns that predict long-term success. Success is defined as users remaining active for more than 30 days. The dataset was processed to create user level features such as total activity, tool usage and early engagement metrics. A Random Forest model was used to evaluate feature importance. The results show that early user activity and overall engagement are the strongest predictors of success, while credit usage has minimal impact. These insights can help improve onboarding and retention strategies.