This project develops a machine learning model to predict customer churn using Python. It includes data analysis, feature engineering, and model training with scikit-learn, optimized using GridSearchCV across multiple models. The final model is deployed as an interactive Streamlit web app for real-time churn predictions. This project is designed for beginners and professionals to learn churn prediction, gain insights, and showcase results through a user-friendly interface.
- Data Analysis: Visualize and explore customer data to uncover churn patterns.
- Feature Engineering: Process and create features to improve model performance.
- Machine Learning: Train and optimize multiple models (4–5) using scikit-learn with GridSearchCV for hyperparameter tuning.
- Streamlit App: Deploy an interactive web app to predict churn based on user inputs.
- Dataset: Expects a churn dataset (https://www.kaggle.com/datasets/abdullah0a/telecom-customer-churn-insights-for-analysis) with features like tenure, charges, and a binary churn column.
- Python Libraries:
- pandas, numpy: Data processing.
- matplotlib, seaborn: Data visualization.
- scikit-learn: Machine learning and GridSearchCV.
- streamlit: Web app deployment.
- jupyter: Notebook environment.
- Install dependencies via
requirements.txt:pandas>=2.0.0 numpy>=1.24.0 matplotlib>=3.7.0 seaborn>=0.12.0 scikit-learn>=1.2.0 streamlit>=1.20.0 jupyter>=1.0.0
notebooks/churn_prediction.ipynb: Notebook for data analysis, feature engineering, and model training.app.py: Streamlit app for deploying the churn prediction model.data/churn_data.csv: Sample or user-provided dataset.requirements.txt: List of dependencies.
The app is deployed on Streamlit Community Cloud. Access it at https://customerchurnpredictionapp-eqfjc6rniyydkr5xkayhze.streamlit.app/
Build and deploy your churn prediction model with ease! 🚀
For any help, feel free to reach out to me 🙂
- Email: Drop a Mail
- LinkedIn: Connect with me
- Instagram: Follow me