Formula 1 Prediction for the Japanese Grand Prix 2025

This project predicts the winner and podium finishers for the 2025 Japanese Grand Prix at Suzuka using historical Formula 1 (F1) data from 1950 to 2024.The prediction model is an ensemble of XGBoost and LightGBM classifiers, trained on a dataset of past Suzuka races. The project also evaluates the model's performance using various error metrics, such as accuracy, confusion matrix, log loss, ROC-AUC score, and mean absolute error (MAE) for predicted probabilities.

The 2025 season includes significant driver and team changes, which are incorporated into the prediction. The final prediction considers factors like weather forecasts (dry, 21°C) and the latest team dynamics.

Project Structure

The project is organized as follows:

• predict_2025_japanese_gp.ipynb: The main Jupyter Notebook containing the code for data loading, preprocessing, feature engineering, model training, and prediction.
• data/: This directory should contain the CSV files with the F1 data (races.csv, results.csv, drivers.csv, constructors.csv, qualifying.csv, lap_times.csv, pit_stops.csv).

Data

The project uses historical Formula 1 data from the Ergast Developer API, which can be found on Kaggle: Formula 1 World Championship (1950-2024). Make sure to download the data and place it in the data/ directory before running the code.

Dependencies

Install the required Python packages using pip :

pip install pandas numpy matplotlib seaborn scikit-learn xgboost lightgbm imbalanced-learn

Methodology

1. Data Loading and Preprocessing: The code loads the data from the CSV files and merges relevant information into a single DataFrame.
2. Feature Engineering: Several features are engineered to improve model accuracy, including driver form, constructor form, Suzuka-specific performance, driver age, qualifying time difference, pole position indicator, Suzuka race experience, recent form, fastest lap potential, tire strategy, and weather impact.
3. Model Training: An ensemble model is trained using XGBoost and LightGBM classifiers. SMOTE is used to handle class imbalance in the dataset. Hyperparameter tuning is performed using GridSearchCV to optimize model performance.
4. Prediction: The trained model is used to predict the winner and podium for the 2025 Japanese Grand Prix based on input data.

Usage

1. Make sure you have the necessary libraries installed (pandas, numpy, scikit-learn, xgboost, lightgbm, imblearn).
2. Place the F1 data files in the data/ directory.
3. Open the predict_2025_japanese_gp.ipynb notebook in Google Colab or Jupyter Notebook.
4. Run the notebook cells sequentially.

Results

The model's performance is evaluated using metrics such as accuracy, classification report, log loss, ROC-AUC score, and mean absolute error. The prediction results for the 2025 Japanese Grand Prix are displayed in the notebook, showing the predicted winner and podium.

Future Improvements

• Incorporate more advanced weather data.
• Add more relevant features, such as tire compound and pit stop strategy.
• Explore other machine learning models.
• Improve hyperparameter tuning and model evaluation techniques.