🏎️ Formula 1 Race Predictor

A machine learning project that predicts the finishing order of all 20 drivers in a Formula 1 race based on historical data, driver stats, and current grid positions. Built using FastF1, Scikit-Learn, XGBoost, and Streamlit.

---

📦 Features

• Predict full race finishing positions
• Supports all 2024 races (auto-loaded from FastF1)
• Inputs: Grid position per driver
• Feature engineered with:
  • Average grid & race positions
  • Driver & team experience
  • Total Points from 2024
  • Qualifying score
• Deployable via Streamlit Cloud

---

🚀 How It Works

1. Model

• Trained with: StackingRegressor
• Base estimators: Ridge, Lasso, XGBoost, GradientBoosting
• Final estimator: Ridge or Lasso
• Feature scaling: StandardScaler
• Saved as: stack_model.pkl, scaler.pkl, feature_columns.pkl

2. Data

• Source: FastF1
• Final dataset: f1_final_data.csv, merged with driver/team stats
• Manually created files:
  • filtered_drivers_info.csv

3. Streamlit App

• UI includes:
  • Dropdown to select race
  • Input grid positions for all 20 drivers
  • Display predicted order

---

🛠️ Setup

🔗 Requirements

pip install -r requirements.txt

📂 Folder Structure

Formula1-Predictor/
├── fastf1_guide.md               # Guides to all commands in fastf1
├── app.py                        # Streamlit app
├── collect_f1_data.ipynb         # Collects the data from fastf1
├── build_features.ipynb          # Making the final dataset for the model
├── train_model.ipynb             # Main training file of the model
├── model/
│   ├── f1_race_predictor_model.pkl        # Trained model
│   ├── scaler.pkl                         # Scaler
│   └── feature_columns.pkl                # Column order
├── DATA/
│   ├── f1_results_2024_2025.csv           # Combined race data
│   ├── f1_drivers_points_exp.xlsx         # Driver's Experience
│   ├── f1_final_data.csv                  # Final dataset for model training
│   └── filtered_drivers_info.csv          # Final dataset for driver's information
└── requirements.txt

---

🌐 Live App

You can try the live app here:
https://formula1-predcition-model.streamlit.app/

---

🌍 Deployment

1. Push this project to GitHub
2. Go to streamlit.io/cloud
3. Connect repo, select app.py
4. Click Deploy

---

📊 Usage

streamlit run app.py

• Select a race (2024 calendar)
• Enter grid positions for each driver
• Click Predict to see results

---

🧠 Author

Built by Nitya Mehta using FastF1, Scikit-Learn, XGBoost, and pure Formula 1 passion.

Built With

• git
• github
• numpy
• pandas
• python
• scikit-learn
• seaborn
• streamlit
• xgboost