🏎️ Formula 1 Predictor

Welcome to our Formula 1 Predictor project! This project was made for the Applied Machine Learning course at the University of Groningen, part of the Artificial Intelligence Bachelor's program.

The goal is to predict the outcome of Formula 1 races leveraging machine learning.

Getting Started

Currently, our API is not publicly available, but you can still run the code locally to train and test our diverse selection of models.

All changeable parameters are stored in the config.py file, so you can easily adjust them to your needs.

Running the Project Locally

[WARNING] This project is built with Python 3.9. Make sure you have this version installed on your machine. If you are using a different version, you may encounter compatibility issues.

1. Clone this repository to your local machine:

   git clone https://github.com/1-million-weed/Applied-ML-GROUP1.git
   cd Applied-ML-GROUP1

2. Set up a virtual environment (our team used a mix between pipenv and venv):

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

OR

2. If you prefer using pipenv, you can install the dependencies with:

   pip install pipenv
   pipenv install

   Ensure you are in the directory where the Pipfile is located when running this.
3. To run just the API server, execute:

   python main.py

   Then head over to 'http://localhost:8000/docs' to access the Swagger UI and test the API endpoints.
4. To run the Streamlit application, execute:

   streamlit run main.py

   Then head over to 'http://localhost:8501' to acces our streamlit api demo.

THE CONFIG FILE

Everything in this project is configurable through the config.py file. This includes:

• The model to be used for predictions
• Dataset acquisition
  • Whether to acquire new data or use existing data
  • Whether to preprocess the data
  • Whether to generate new features from raw data
  • Training and testing data split
• Model training
  • Whether to train a new model or use an existing one
  • Tensorboard logging
• Model evaluation
  • Whether to evaluate the model on the test set
  • Whether to show the evaluation plots
• Model inference
  • Whether to run inference on the model
  • Whether to activate the API server
  • Whether to run the Streamlit application
• Logging
  • The level of logging (DEBUG, INFO, WARNING, ERROR, CRITICAL)

API Endpoints

The API provides several endpoints to interact with the model:

• GET /: API Health check

• GET /meetings: List of all meetings from the current season

• GET /meetings/{meeting_id}/max-laps: List the maximum amount of laps in a race

• GET /docs: Swagger UI documentation

• POST /predict: Predict the outcome

Notable references:

• Elo calculation code

Project Structure

├───data  # Stores raw .csv files
├───models  # Stores .pkl files for trained models
├───experimental  # Contains experimental .ipynbs & .py
├───f1_predictor
│   ├───app # Contains the Streamlit app
│   ├───data  # stores processed .csv files
│   ├───data_acquisition # For acquiring data from the FastF1 API for 2025 data
│   ├───features # For scripts and logic for feature engineering
│   ├───ml # Contains the machine learning logic (pipelines & managers)
│   └───models  # For model creation, not storing .pkl
├───reports # For outputs and visualisations
├───tests
│   ├───data
│   ├───features
│   └───models
├───.dockerignore
├───.gitignore
├───.pre-commit-config.yaml
├───config.py
├───Dockerfile
├───main.py
├───mylogger.py
├───train_model.py
├───Pipfile
├───Pipfile.lock
├───README.md
├───requirements.txt