Customer-Churn-Prediction

What the project does – Predicts whether a telecom customer will churn using their demographic, account, and service usage data.

Key features – Data cleaning (numeric conversion), exploratory visualizations, multiple classification models (Logistic Regression, Decision Tree, Random Forest, SVM, XGBoost), comprehensive evaluation (accuracy, precision, recall, F1, ROC‑AUC, confusion matrix), and model selection (Random Forest ~85% accuracy).

Tech stack – Python, Pandas & NumPy, Matplotlib & Seaborn, Scikit‑learn, XGBoost, Jupyter Notebook.

Use case – Enables subscription‑based businesses to identify at‑risk customers early and implement retention strategies, reducing revenue loss from churn.