SkyPredict is a machine learning-based system designed to predict domestic flight ticket prices in India using real-time scraped data and a suite of regression algorithms. This project aims to bring transparency to dynamic airfare pricing and help users make informed ticket booking decisions.
- Collected over 65,000 flight records using Python-based web scraping from MakeMyTrip and EaseMyTrip.
- Cleaned and preprocessed data to handle nulls, outliers, and encoded categorical variables.
- Trained and evaluated 10+ regression models including:
- Random Forest Regressor
- K-Nearest Neighbors (KNN)
- Decision Tree
- XGBoost
- Bagging & Extra Trees Regressor
- Achieved:
- R² Score: 0.76 (Random Forest)
- MAPE: 12.7%
- RMSE: ~2779
- Evaluated using metrics: R², MAE, RMSE, MAPE