Orange_Telecom-Customer-churn-prediction

A supervised machine learning project aimed at predicting customer churn for Orange Telecom using real-world telecom usage data. This project analyzes customer behavior and service interactions to identify key churn indicators and build predictive models to support targeted retention strategies.

Project Structure

• Raw dataset (churn-bigml-80.csv)

• Jupyter notebooks for EDA, modeling, and tuning

• Report

Problem Statement

Objective: Predict whether a telecom customer will churn based on demographic information, service plans, and usage metrics. This will allow Orange Telecom to proactively retain customers and reduce revenue loss.

Technologies & Tools

• Python (pandas, numpy, scikit-learn, seaborn, matplotlib)

• Jupyter Notebook

• GridSearchCV (Hyperparameter tuning)

• Feature Engineering & Importance Analysis

Models Used

• Logistic Regression (Baseline)

• Decision Tree

• Random Forest

• Gradient Boosting (Best performing model)