🚀 Customer Churn Prediction API (Microservice)

📌 Project Overview

This project transforms a trained Machine Learning model into a professional, scalable web service. By wrapping the model in a FastAPI framework and containerizing it with Docker, we ensure the model works consistently across any environment.

🛠️ Tech Stack

• Languages: Python
• Libraries: FastAPI, Pydantic, Scikit-Learn, Pandas, Uvicorn
• DevOps: Docker

🏗️ Architecture

• Backend: FastAPI (Python)
• Model: Random Forest Classifier (Previous Project logic)
• Containerization: Docker (Python 3.10-slim)
• Documentation: Automatic OpenAPI/Swagger UI

🧠 Model Origin

The churn_model.pkl used in this microservice was developed in my previous Churn Analysis Project. In that project, I performed extensive Exploratory Data Analysis (EDA) and trained a Random Forest Classifier to identify patterns in customer behavior. This project focuses on the deployment phase of that model's lifecycle.

🧪 Live API Demonstration

Below is a real-time prediction captured from the interactive Swagger UI. The API successfully processes a "passenger" profile and returns a churn status in milliseconds.

🚀 How to Run Locally

1. Build the Docker Image

docker build -t churn-api .

2. Launch the Container

docker run -p 8000:8000 churn-api

3. Test the Endpoint

Navigate to http://localhost:8000/docs in your browser. Use the POST /predict endpoint.