Skip to content
← All projects
Data Science / AnalyticsCompleted

Value-Weighted Telecom Churn Retention

An end-to-end churn-retention proposal that joins 100K telecom customers with usage records, predicts churn risk, and converts model scores into a value-weighted targeting strategy.

Overview

A business-facing churn analytics project that turns model scores into a targeting strategy, lift analysis, and estimated financial impact.

Written for analytics and DS roles: model quality matters, but the strongest story is value-weighted decision-making.

Impact

  • Selected XGBoost at ~0.70 ROC AUC after benchmarking Logistic Regression, Random Forest, and XGBoost.
  • Top risk decile showed 79.5% churn; targeting the top 20% captures ~30% of churners.
  • Projected $2.07M/yr net benefit and 2.1x ROI under stated retention-offer assumptions (modeled, not measured).

What it shows

  • Churn modeling, decile lift, cumulative gains, and sensitivity analysis.
  • Business translation from model score to retention-offer strategy.
  • A measurable ROI story - stated clearly as a projection from assumptions, not an achieved outcome.

Demo plan

  • Embedded dashboard or Streamlit page with decile targeting and ROI assumptions.
  • Add repo or notebook after removing local paths and making the analysis reproducible.

Stack