Customer Churn Prediction
Interpretable churn model using decision trees — feature engineering, evaluation, and explainable split logic for actionable retention insights.
Overview
A churn prediction project built around decision trees for interpretability. The emphasis is on creating a model that both predicts and explains *why* customers leave.
Problem
Churn models are only useful if teams can act on them. Black‑box accuracy without explanations often fails to translate into retention strategy.
Approach
Prepared the dataset with thoughtful preprocessing and feature engineering. Trained and tuned decision‑tree models, then analyzed splits to surface drivers behind churn. Evaluated performance with standard metrics while prioritizing interpretability and actionable outputs.
Impact
An interview‑friendly ML case study: it demonstrates modeling fundamentals and the product mindset of making predictions explainable and usable.