This project aims to create a point scoring model for Fantasy Premier League (FPL) players using a Poisson regression approach. The FPL is a popular fantasy sports game where participants build virtual teams of real players from the English Premier League and earn points based on their performance in actual matches.
The project utilizes data from the Fantasy Premier League GitHub repository (specifically, the data/ directory) as the primary data source. This data contains comprehensive information about player statistics, fixtures, and historical FPL scores.
The core methodology employed in this project is Poisson regression. Poisson regression is a statistical technique commonly used for modeling count data, making it suitable for modeling player point scoring in fantasy sports, where points are awarded for goals, assists, clean sheets, and other events.
There are 3 Jupyter notebooks data_cleanin.ipynb, data_exploration.ipynb and data_model.ipynb which should be easy to follow along.