Flight-Price-Predictor

A machine learning model which predicts the prices of domestic flights to and from any of the metro cities based on various parameters like airline, number of stops/layovers, arrival and departure times, origin, destination, ticket class, and the number of days from the trip.

Overview

This project implements a prediction system which can be used to predict the prices of flights based on various parameters. The machine leaning model analyses historical data to estimate flight prices.

Features

• Input Parameters: airline, number of stops/layovers, arrival and departure times, origin, destination, ticket class, and the number of days from the trip.
• Model: Random Forest (Regressor)
• Price Estimation: Estimates the cost of domestic flights.
• Streamlit Frontend: Simple frontend built using Streamlit.

Dataset

The dataset used has been taken from Kaggle.

• airline: The name of the airline company.
• flight: The flight code of the plane.
• source_city: The city from which the flight takes off.
• departure_time: The departure time of the flight.
• stops: The number of stops the flight makes en-route.
• arrival_time: The arrival time of the flight.
• destination_city: The city where the flight will land.
• class: The ticket class (Economy or Business).
• duration: The time required to travel between the two cities.
• days_left: The number of days left for the flight/trip.
• price: The ticket price (in INR).

Installation

1. Clone the repository:

git clone https://github.com/sayanjit082805/Flight-Price-Predictor.git
cd Flight-Price-Predictor

2. Create a virtual environment (recommended):

python -m venv flight_predictor_env
source flight_predictor_env/bin/activate  # On Windows: flight_predictor_env\Scripts\activate

3. Install required dependencies:

pip install -r requirements.txt

Usage

Command Line/Locally

A sample test.py has been provided, simply tweak the required parameters and run the file using python3 test.py.

# features = ['airline', 'stops', 'source_city', 'departure_time', 'arrival_time', 'destination_city', 'class', 'days_left']
import joblib
import pandas as pd
model = joblib.load("flight_price_model.pkl")
model_columns = joblib.load("flight_price_model_columns.pkl")
new_price = pd.DataFrame(
    [
        {
            "airline": "Air_India",
            "stops": "zero",
            "source_city": "Kolkata",
            "departure_time": "Morning",
            "arrival_time": "Afternoon",
            "destination_city": "Mumbai",
            "class": "Economy",
            "days_left": 5,
        }
    ]
)
new_price_encoded = pd.get_dummies(new_price)
new_price_encoded = new_price_encoded.reindex(columns=model_columns, fill_value=0)
predicted_price = model.predict(new_price_encoded)
print(f"The price of the ticket is: {round(predicted_price[0]):,} INR")

Sample Output

The price of the ticket is: 8,191 INR

Deployment

Deployed here.

Model Performance

Data	MAE	R² Score	MSE
Training Data	2,333	0.965	18234775.38
Test Data	2,344	0.964	18326984.04

Acknowledgements

The dataset has been taken from kaggle.

License

This project is licensed under The Unlicense License.