Data analytics applied to football and football players

Abstract

Football is the most popular sport in the world and it is undergoing a process of implementing analytics in decision making. This quantitative study aims to identify a set of Technical Performance Indicators for players of 7 different positions and analyse the aggregate performance of the players of those positions. The positions used were: Goalkeeper, Full Back, Centre Back, Holding Midfielder, Attacking Midfielder, Wide Midfielder and Striker.
To be able to conduct these analyses, the match data of the technical actions of the Portuguese Primeira Liga 2016/2017 season players were extracted from an online platform and an Access database was developed.
The identification of the indicators was successfully achieved by a clustering analysis, which used the K-Means technique, and a quantile analysis. These studies revealed 16 Technical Performance Indicators that varied and had different values of relevance from position to position.
Afterwards, a Data Envelopment Analysis (DEA) model was used to compile the technical indicators and analyse the aggregated performance of the players using that composite indicator. This last study presented results, however, due to limitations related to the weight awarded to the outputs, they are not the most accurate.
The present research study was based on the academic article from Hughes, et al. (2012), “Moneyball and soccer - an analysis of the key performance indicators of elite male soccer players by position” and it tried to prove analytically what was only empirically stated in literature before.

Date of Award	11 Jul 2018
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Maria Silva (Supervisor)