01 / The brief
The question
Build a churn prediction model that helps identify telecom customers at risk of leaving.
This GitHub project predicts telecom customer churn using machine learning. It is relevant for the portfolio because churn prediction is a strong business analytics use case for customer retention.
Results & observations
- Created a churn prediction workflow for telecom customer retention.
- The project demonstrates applied machine learning, classification, and customer analytics.
Challenges
- Preparing customer account data
- Handling mixed categorical and numerical features
- Evaluating churn classification performance
- Making churn insights understandable
Approach
- Cleaned and prepared telecom customer data.
- Processed contract, billing, and service features.
- Trained classification models for churn prediction.
- Evaluated model performance and churn risk indicators.
Implementation details
Data Cleaning
Cleaned telecom customer records.
Feature Preparation
Prepared contract, billing, and service features.
Model Training
Trained machine learning models to predict churn.
Evaluation
Measured model performance using classification metrics.
Retention Insight
Connected model output to retention decision-making.
Explore the report
Select a screenshot for a larger view.