Flight Delay Prediction API

A machine learning web application that predicts flight delays with two frontend options: FastAPI with HTML templates and Streamlit.

Project Structure

Flight_Delay_Predictor/
├── main.py                 # FastAPI application with HTML frontend
├── api.py                  # Streamlit application
├── model/                  # Trained ML models
│   ├── flight_delay_model.pkl
│   └── model_columns.pkl
├── templates/              # HTML templates for FastAPI
│   └── index.html         # FastAPI frontend interface
├── venv/                   # Virtual environment
├── requirements.txt        # Python dependencies
├── start_fastapi.bat      # FastAPI startup script
├── start_streamlit.bat    # Streamlit startup script
└── test_api.py            # Test script

Setup

1. Virtual environment is already created and activated
2. Dependencies are already installed

Running the Application

Option 1: FastAPI Frontend (HTML/JavaScript)

1. Start the FastAPI server:

   start_fastapi.bat

   Or manually:

   venv\Scripts\activate
   uvicorn main:app --reload

2. Access the application:

   • Web Interface: http://localhost:8000
   • API Documentation: http://localhost:8000/docs
   • Alternative API Docs: http://localhost:8000/redoc

Option 2: Streamlit Frontend (Python-based UI)

1. Start the Streamlit app:

   start_streamlit.bat

   Or manually:

   venv\Scripts\activate
   streamlit run api.py

2. Access the application:

   • Streamlit Interface: http://localhost:8501

API Endpoints (FastAPI)

GET /

Serves the HTML frontend interface for making predictions.

POST /predict

Makes a flight delay prediction.

Request Body:

{
  "month": 1,
  "day_of_month": 15,
  "day_of_week": 3,
  "op_unique_carrier": "WN",
  "origin": "LAX",
  "dest": "SFO",
  "crs_dep_time": 1400,
  "dep_delay": -5.0,
  "distance": 337.0
}

Response:

{
  "prediction": 0,
  "prediction_label": "On Time",
  "confidence_on_time": "0.8523",
  "confidence_delayed": "0.1477"
}

Features

FastAPI Frontend

• Web Interface: Clean HTML form with real-time predictions
• REST API: JSON-based API for programmatic access
• Interactive Documentation: Auto-generated API docs with Swagger UI
• Responsive Design: Works on desktop and mobile devices

Streamlit Frontend

• Interactive UI: Python-based web interface with sliders and inputs
• Real-time Feedback: Instant predictions with confidence metrics
• Visual Elements: Progress bars and colored result displays
• Form Validation: Built-in input validation and error handling

Model Information

The model uses the following features:

• Month (1-12)
• Day of Month (1-31)
• Day of Week (1-7)
• Carrier Code (e.g., WN, DL, AA)
• Origin Airport (e.g., LAX)
• Destination Airport (e.g., SFO)
• Scheduled Departure Time (HHMM format)
• Departure Delay (minutes, negative for early)
• Distance (miles)

Testing

Run the test script to verify the setup:

python test_api.py

Repository

This project is available at: https://github.com/ChukwuemekaP1/figth_predict_model.git