Challenge 11 Football Profiling for Player Performance Analysis 

Challenge Description

Good player profiles are essential when it comes to making the right tactical decisions about playing positions. For this, player data from the top five European leagues was scraped for the purpose of creating such profiles and capturing the characteristics of the players in each position. This will help FC Servette not only to fill its positions optimally but can also used to scout for players. FC Servette has not won the Swiss Super League title since 1999, and your work thus will make a difference in winning the next one. So go for it! 

Challenge Owner

FC Servette is a Swiss Super League Club looking to improve the performance of its teams.

Solution

Pitch

The challenge aims to help FC Servette create profiles of its players. This would allow it to gain an overview of its athletes’ performance, plan its scouting activities and better develop players’ potential based on the target profile. The profiling was based on data from the five European leagues (Sources: Statbombs, FBREF) and the Super League, and it includes player statistics such as the number of shots, key passes, percentage of successful dribbles, among other things. 

The challenge team explored and analyzed the data to identify its salient characteristics, including outliers. Based on a knowledge of football, the variables were grouped into performance clusters, each showing the profile of the player. For example, the Striker role has six clusters: Scoring chance generator, Target Man, Finisher, Selfish & Risky, Dribblers, and Efficient Attacking Creator. The Servette strikers can then be assigned to these clusters.

The heatmap the figure below shows the "importance" of each performance variable for the given cluster on a scale from 0 to 1. 

This approach to data analysis and visualisation also helps to compare the clusters in terms of their key characteristics:

Potential next steps

• Select features by using machine-learning algorithms such as linear regression 
• Use other clustering methods to see the changes in clusters
• Consider bias when integrating information about changes in the team’s strategy and coaching approach and observing the changes in the results
• Include event data to understand the circumstances under which the player performed in a certain way

Data