Player Recommendation System for Fantasy Premier League

Abstract

Fantasy Premier League turns sports fandom into a recurring decision-making problem, where users must balance player form, price, and team loyalty under uncertainty.

This project built a player recommendation system using FPL API data, visualizations, and statistical analysis to evaluate players with richer context than simple Return on Investment alone.

The core goal was to reduce favoritism bias, generate more objective recommendations, and turn season data from the 2021-22 Premier League into an interpretable decision-support workflow.

What I Built

• FPL API data supported a recommendation workflow grounded in player form, team context, and comparative performance metrics.
• ROI-only evaluation proved too narrow for realistic player selection because it ignored fan bias and richer performance signals.

Impact

• Turned raw match and player data into a decision-support system that helps users make less emotional picks.
• Produced a publishable research output presented at a conference and archived on IEEE Xplore.