alaa-m0hammed/Customer-Churn-Prediction_DEPI ? reverse-engineered prompt

Build me a customer churn prediction app from this repo that feels ready to demo and deploy.

I want it to take telecom customer details, predict the chance they will churn, and show a simple risk label like low, medium, or high with a clear recommendation. There should be an API I can call from a form or another app, plus a small dashboard where I can see the main churn insights, model performance, and make live predictions. Please use the trained model and saved feature files in the project, and make sure the input fields match the data the model expects.

Also include the monitoring and retraining pieces already described here, so the project can track experiments, detect drift, and support automatic retraining when the data changes. Keep it easy to run locally, and make the Docker setup work too. If anything is unclear, check the existing notebooks and README first, and look up current docs online if you need to.