Flight Fare Predictor

This repository contains an end-to-end Flight Fare Prediction application.The project demonstrates the practical application of Machine Learning, REST APIs, and a modern frontend to solve a real-world regression problem.

Project Overview

The goal of this project is to predict flight ticket prices based on various travel-related parameters such as airline, source, destination, journey date, duration, and number of stops.

The system is divided into three major parts:

• Machine Learning Model
• Backend API
• Frontend UI

Machine Learning Details

• Algorithm Used: Random Forest Regressor
• Problem Type: Regression
• Dataset Size: ~10,600 flight records
• Target Variable: Flight Price

The trained model is serialized using joblib and served through an API.

---

Project Structure

Parent Folder	Sub-Modules	Description
Backend	App	FastAPI-based backend service that exposes REST APIs to interact with the trained flight fare prediction model
	Models	Saved trained model files (.joblib) used by the API for inference
	ML	Contains machine learning code including model training, testing, and prediction logic
	Data	Stores training and test datasets used for building and validating the model
Frontend	React UI	React-based web application that allows users to input flight details and view predicted fares

Technologies Used

Backend & Machine Learning

Frontend

Getting Started

Prerequisites

• Python 3.10+
• Node.js (v16+ recommended)
• npm / yarn / pnpm

1. Clone the Repo

git clone https://github.com/vivek-src/flight-fare-predictor.git
cd flight-fare-predictor

2. Backend Setup

#cd to backend folder
cd backend

#Create a python virtual environment
python3 -m venv venv

#Activate virtual environment
source venv/bin/activate

#Install Python Dependencies
pip install -r requirements.txt

#Train the Machine Learning Model
python3 ml/train.py

#Start the FastAPI Server
uvicorn app.main:app --reload
#server will start at 'http://127.0.0.1:8000', test in browser 'http://127.0.0.1:8000/docs'

The setup process is slightly different for Windows environments

3. Frontend Setup

#Open a new terminal and run
cd frontend
npm install
npm run dev

#The frontend UI will be available at http://localhost:5173

Use the Frontend UI or API endpoint to Use the model

#example payload

{
  "airline": "IndiGo",
  "source": "Delhi",
  "destination": "Cochin",
  "date_of_journey": "2024-06-10",
  "dep_time": "10:00",
  "arrival_time": "12:30",
  "duration": "2h 30m",
  "total_stops": 0,
  "additional_info": "No Info"
}

Screenshot

Model Performance

The Random Forest Regression model was trained on approximately 10,600 flight records.

Evaluation Metrics

• Mean Absolute Error (MAE): 695.47
• Root Mean Square Error (RMSE): 1654.54
• R² Score: 0.87

Limitations

• Flight duration has a higher influence on predictions than some other features.
• Predictions may fail for cities or routes not seen during training.
• Real-time factors such as demand, holidays, and seat availability are not considered.