IEEE VIS 2024 Content: Team-Scouter: Simulative Visual Analytics of Soccer Player Scouting

Team-Scouter: Simulative Visual Analytics of Soccer Player Scouting

Anqi Cao - Zhejiang University, Hangzhou, China

Xiao Xie - Zhejiang University, Hangzhou, China

Runjin Zhang - Zhejiang University, Hangzhou, China

Yuxin Tian - Zhejiang University, Hangzhou, China

Mu Fan - Zhejiang University, Hangzhou, China

Hui Zhang - Zhejiang University, Hangzhou, China

Yingcai Wu - Zhejiang University, Hangzhou, China

Room: Bayshore V

2024-10-17T14:15:00ZGMT-0600Change your timezone on the schedule page
2024-10-17T14:15:00Z
Exemplar figure, described by caption below
System user interface. The interface contains two views: a navigation view (A) and an investigation view (B). The navigation view consists of a squad board (A1) to navigate players will be replaced and a player ranking list (A2) to compare players by personal information and performances. The investigation view includes an on-ball tactic list (B1) for exploring essential on-ball tactics, a player record list (B2) to compare players' simulated actions under a certain on-ball tactic, and a simulated action map (B3) to display players' detailed simulated actions.
Fast forward
Keywords

Soccer Visualization, Player Scouting, Design Study

Abstract

In soccer, player scouting aims to find players suitable for a team to increase the winning chance in future matches. To scout suitable players, coaches and analysts need to consider whether the players will perform well in a new team, which is hard to learn directly from their historical performances. Match simulation methods have been introduced to scout players by estimating their expected contributions to a new team. However, they usually focus on the simulation of match results and hardly support interactive analysis to navigate potential target players and compare them in fine-grained simulated behaviors. In this work, we propose a visual analytics method to assist soccer player scouting based on match simulation. We construct a two-level match simulation framework for estimating both match results and player behaviors when a player comes to a new team. Based on the framework, we develop a visual analytics system, Team-Scouter, to facilitate the simulative-based soccer player scouting process through player navigation, comparison, and investigation. With our system, coaches and analysts can find potential players suitable for the team and compare them on historical and expected performances. For an in-depth investigation of the players' expected performances, the system provides a visual comparison between the simulated behaviors of the player and the actual ones. The usefulness and effectiveness of the system are demonstrated by two case studies on a real-world dataset and an expert interview.