Fantasy Football Team Optimizer

This project utilizes linear regression and linear programming to suggest an optimized fantasy football team. The approach focuses on maximizing long-term performance by leveraging statistical modeling to select players while adhering to budget constraints.

Features

• Linear Regression: A model is trained on important features to predict player scores based on historical data.
• Linear Programming: Pyomo is used to optimize player selection with constraints like budget and team composition.
• Customizable Settings: Users can adjust the budget and the number of players based on their league's specific rules.
• Long-term Focus: This model emphasizes consistent, long-term performance by not considering short-term variables like match difficulty.

Workflow

1. Linear Regression:

   • The model uses important features of players to learn the parameters for predicting performance.
   • A new column, 'score', is added to the dataset to represent the predicted points for each player.

2. Linear Programming (Pyomo):

   • The optimization problem is defined with a budget constraint (initially 83.5 million, after subtracting the value of bench players).
   • The program selects 11 players within the budget that maximize the total score.

3. Customization:

   • To adjust the number of players or the budget, simply modify the variables budget and num_players in the code.
   • For example, if you want to add a bench player, increase the number of players to 12 and adjust the budget accordingly.

Getting Started

1. Download the Premier League Dataset:

   • Get the dataset from Kaggle: Fantasy Premier League Dataset 2024-2025.

2. Upload the Dataset:

   • Upload the dataset to your Google Colab environment.
   • Copy the path to the file and paste it into the read_csv() function in the notebook.

3. Run the Notebook:

   • Open the notebook in Google Colab, execute the cells, and the model will optimize and suggest a team.

Future Improvements

• Supervised Captaining:

  • Develop a strategy to recommend captains based on fixture difficulty and player consistency over a long period.

• Complex Objective Function:

  • Replace linear regression with a more sophisticated model (e.g., deep learning) to better capture player performance.

• Transfers:

  • Introduce a transfer system where players can be swapped based on performance. Prediciting the players next round points based on several variables. Then they could be replaced with better-performing alternatives.
  • Incorporate injury/red card considerations into the model.

• Simplified Premier League Formula:

  • Implement the official scoring formula provided by the Premier League for a simpler approach instead of Machine Learning.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• Thanks to Kaggle for providing the dataset.
• Special thanks to the open-source community for Pyomo and the libraries used.

Images