A machine learning web application that predicts whether a telecom customer will churn (leave) or stay, based on their usage patterns and history.
🔗 Live App: https://churn-prediction-ipwth9dtpfwmu5byjdpyyr.streamlit.app
📂 GitHub: https://github.com/pawan0221/churn-prediction
Telecom companies lose millions every year due to customer churn. This app uses Machine Learning to identify at-risk customers early, so the business can take action before they leave.
- 🔍 Predicts churn in real time based on customer inputs
- 📊 Shows confidence score with every prediction
- 🌲 Powered by Random Forest (100 decision trees)
- 🌐 Fully deployed — no installation needed
| Detail | Info |
|---|---|
| Algorithm | Random Forest Classifier |
| Library | scikit-learn |
| Training Data | 300 synthetic telecom customer records |
| Features Used | Age, Data Usage, Complaints, Tenure, Plan Type |
| Target | Churn (1 = Yes, 0 = No) |
| Feature | Description |
|---|---|
| Age | Customer age (18–70) |
| Data Usage (GB) | Monthly data consumption |
| Number of Complaints | Complaints filed in last 6 months |
| Tenure (months) | How long they've been a customer |
| Plan Type | Prepaid or Postpaid |
| Tool | Purpose |
|---|---|
| Python | Core language |
| pandas & numpy | Data processing |
| scikit-learn | ML model training |
| joblib | Model saving/loading |
| Streamlit | Web app & deployment |
| Git & GitHub | Version control |
# Clone the repo
git clone https://github.com/pawan0221/churn-prediction.git
cd churn-prediction
# Install dependencies
pip install -r requirements.txt
# Run the app
streamlit run app.pyPawan Soni
GitHub: @pawan0221
This project is open source and available under the MIT License.