Customer Inactivity Detection in Banking

This was one of the first complete data science projects I developed independently—from data cleaning to model deployment. Using a Kaggle dataset on bank customer inactivity, I explored behavioral patterns and implemented a predictive model to classify customers at risk of becoming inactive. Beyond its academic value, the project inspired the development of similar analytics strategies within my current role at PBZ Bank, demonstrating how open-source exploration can translate into concrete business innovation.

Key Contributions:

• Cleaned and processed raw banking data using Pandas

• Performed exploratory data analysis and visualizations (churn patterns, inactivity rates)

• Engineered features and tested multiple classification models (Logistic Regression, Random Forest, XGBoost)

• Evaluated models using accuracy, precision, recall, and ROC-AUC metrics

• Reflected insights back into real-world context within PBZ internal analytics