Customer Churn Prediction with Machine Learning

A machine learning project that predicts which telecom customers are likely to leave, using data from 7,043 customers.

• Data cleaning, visual analysis (EDA) and outlier detection

• Feature encoding and selection of the top 25 features

• Built and compared Logistic Regression and Decision Tree models

• Best model: Decision Tree, 80% accuracy, F1 = 0.61

• Main churn drivers: month-to-month contracts, short tenure, fiber-optic internet

Deliverable: a step-by-step Jupyter Notebook with charts and a model comparison.