Flight Price Prediction App is a production-ready machine learning web application that predicts flight ticket prices based on multiple features such as airline, source, destination, stops, and travel time.
This project demonstrates the complete ML lifecycle — from data preprocessing and feature engineering to model optimization and deployment with Streamlit.
💡 Built by Batuhan Başoda — using the [Kaggle Flight Price Dataset] and deployed as a modern ML web app.
- 🧩 Data Preprocessing: Missing value handling, encoding, and feature extraction
- ⚙️ Feature Engineering: Time-based and categorical transformations
- 🧠 Modeling: Trained and compared multiple regression models (RandomForest, XGBoost, etc.)
- 🎯 Hyperparameter Tuning: Used RandomizedSearchCV for best model performance
- 📊 Real-time Predictions: User inputs flight details → instant price prediction
- 🌐 Streamlit UI: Deployed as an interactive, fast, and responsive web app
| Category | Technology |
|---|---|
| Language | Python & Jupyter Notebook & Anaconda Environment |
| ML Libraries | Scikit-learn, XGBoost, NumPy, Pandas |
| Visualization | Matplotlib, Seaborn |
| Web Framework | Streamlit |
| Deployment | Streamlit Cloud |
| Dataset | Kaggle Flight Price Dataset |
🔗 Live Demo: https://your-streamlit-app-link.streamlit.app
- Data split: 80% training / 20% testing
- Model evaluation metrics:
- R² Score: 0.93
- MAE: 1,809
- RMSE: 3,913
- Final model used: Decision Tree Regressor (tuned with RandomizedSearchCV)
# 1️⃣ Clone this repo
git clone https://github.com/batuhanbasoda/flight-price-predictor.git](https://github.com/Batuhan-METU/FlightPricePredictor-MachineLearningModel
cd flight-price-predictor
# 2️⃣ Install dependencies
pip install -r requirements.txt
# 3️⃣ Run the app
streamlit run app.py