NYC Ridge Prediction Model

This project is a machine learning application designed to predict trip durations in New York City using a Ridge regression model. The application includes a FastAPI backend for model training and prediction, and a Streamlit frontend for user interaction.

Setup

Prerequisites

• Docker
• Docker Compose
• Python 3.13

Installation

1. Configure.env.local file variables.

2. Build and run the Docker containers:

   make all

Usage

API Endpoints

• Root: GET /
  • Returns a welcome message.
• Ping: GET /ping
  • Returns a pong message to check API health.
• Predict: POST /predict
  • Accepts a JSON payload with trip details and returns the predicted trip duration.
• Train: GET /train
  • Triggers the model training process and saves the trained model.

Streamlit Frontend

1. Access the Streamlit frontend at http://localhost:8501.
2. Use the sliders to input trip details and click "Predict" to get the trip duration prediction.
3. Click "Train Model" to trigger the model training process.

License

This project is licensed under the MIT License.