Zeeshan Jamal

Built a full-stack churn prediction and revenue intelligence system using the IBM Telco dataset, designed to move beyond model accuracy and quantify real financial impact.

Developed and validated multiple ML models (Logistic Regression, Decision Tree, Random Forest) using 5-fold cross-validation, selecting Logistic Regression as the champion model for production deployment.

The system scores individual customers with churn probabilities and translates predictions into actionable financial metrics, including Monthly Recurring Revenue (MRR) and Annual Recurring Revenue (ARR) at risk.

Key capabilities include:

• Dynamic churn risk segmentation (High / Medium / Low) with adjustable thresholds

• Revenue-at-risk quantification per customer and portfolio level

• Retention ROI simulation (scenario-based churn reduction impact analysis)

• Churn driver exploration with feature-level probability shifts and quartile statistics

• Executive-ready automated PDF reporting

Deployed as a production-style pipeline with:

• End-to-end Jupyter workflow (data processing → modeling → evaluation → reporting)

• Interactive Streamlit dashboard (4 analytical modules)

• Saved, reusable ML model pipeline

• Executive narrative generation for stakeholder communication

Tech Stack: Python, scikit-learn, Pandas, Streamlit, Plotly, ReportLab