Special Section on SIBGRAPI 2017Visual soccer match analysis using spatiotemporal positions of players
Graphical abstract
Introduction
Soccer is a very passionate team sport that attracts the attention of many fans in the world. Soccer has an important economic aspect since billions of dollars are spent to construct successful teams that win titles. However, forming winning teams is not a trivial matter, and soccer managers are looking for any information and analysis tools that can support decisions that impact the success of a team. One example is the extraction of data from the video footage of soccer matches, such as the position of players during the match, possession of the ball, actions performed (e.g., shots, crosses, tackles, fouls, etc.). This data is not easy to evaluate, due to the high volume of information and the high degree of interrelation.
Recent approaches to the scientific analysis of soccer footage data showed success to assist coaches in their decisions on team strategy, opponent analysis, and scouting prospection of players [1], [2]. Most of the approaches described today rely on statistics that span the entire soccer match (e.g., heatmap of a given player on the pitch, possession time, etc.). Only recently we are seeing more sophisticated approaches for this analysis, such as the analysis of soccer player trajectories presented by Shao et al. [3]. We agree that the position of players on the field play an important role in this analysis, and describe alternate ways to visually present this information.
Our proposal relies on two basic visual designs. The first one is the pixel-oriented display proposed by Keim et al. [4]. We use this approach to create a Player Attribute Heatmap (PAH), which is a matrix display of the positions of players during a portion of the match (often a half-time). This visual display allows creating an image that tells a story of how the match developed itself from the perspective of the position of the players in the field. Our PAH can also be parameterized to use different ordering strategies while showing this information (e.g., ordering by vertical or horizontal positions of players), which can review trends such as preferred sides of the field at specific time intervals of the match. We also apply this visual design to display preferred tactical schemes used throughout the match in a Tactical Scheme Heatmap (TSH).
The second approach relies on the pathline glyphs [5] designed for the visualization of unsteady 2D flow. We address with this design the visual clutter of trajectories of players and their inherent complexity. This technique allows creating miniature glyphs that represent player trajectories starting from any part of the field. For example, it is possible to identify for an attacker the preferred directions of movement when starting from a central part of the offensive field.
In summary, the contributions of this paper include:
- •
a system for the visual analysis of soccer matches based on the spatiotemporal position of players;
- •
a parametrizable matrix-oriented design that allows displaying the position of players and tactical schemes throughout the match;
- •
a visual summarization using pathline glyphs of the player trajectories that started from a given region in the field;
- •
an evaluation study of the visual design and feedback from soccer experts.
Fig. 1 gives an overview of the main components of the system, including additional functionality. The data used in this work was generated by our soccer expert collaborators.
Section snippets
Related work
Visualization of sports data is an active and comparably young field of research [6]. MatchPad [7] presents a visual timeline with glyphs to analyze the performance of players in a rugby match, while Chung et al. [8] proposed a system to visualize and interact with sports data in general. There are several systems designed for the visualization of soccer data, and we refer to Gudmundsson et al. [9] for a comprehensive survey of these aspects, and to Anderson and Sally [10] for a historical
Data representation and analysis requirements
In this section, we describe the data format captured from soccer matches. Also, we list a set of questions used as guidelines for the development of our approach and its evaluation. We formulated these questions with the collaborator soccer experts that provided the data and helped evaluate our proposal.
Spatiotemporal visual analysis of players
In this section, we describe the visual designs to analyze the spatiotemporal position of players according to the requirements R1–R6 formulated in the previous section
Results
In this section, we present the main results obtained using our approach and evaluate the visual designs proposed.
Evaluation
We perform two different evaluations of the system. The first evaluation comprised of a demonstration of the system to soccer experts, which gave us feedback on the visual designs proposed and the situations that they might be used for. The second evaluation comprised of an evaluation study where specific tasks were given to subjects, in their majority not soccer experts. We detail the evaluation results obtained below.
Conclusion and future work
In this work, we described visual designs to support the analysis of the spatiotemporal positions of players during a soccer match. The team behind the design of this work included soccer experts, described before, that supplied the soccer data used in the analysis, as well as helped in all stages of development and evaluation. Central to our ideas was the analysis of the evolution of the match from a team or individual perspective. We proposed PAH, a heatmap-based approach to summarize player
Acknowledgments
The authors wish to thank the anonymous reviewers for their valuable comments and suggestions to improve the quality of the paper. This work was supported in part by CNPq process 308851/2015-3 and FAPESP process 2016/05205-1.
References (33)
- et al.
Spatial dynamics of team sports exposed by Voronoi diagrams
Hum Mov Sci
(2012) - OptaPro. http://optasportspro.com; 2017 [accessed...
- Stats. http://www.stats.com; 2017 [accessed...
- et al.
Visual-interactive search for soccer trajectories to identify interesting game situations
Electron Imaging
(2016) Designing pixel-oriented visualization techniques: theory and applications
IEEE Trans Vis Comput Graph
(2000)- et al.
Pathline glyphs
Comput Graph Forum
(2014) - et al.
Towards classifying visualization in team sports
Proceedings of international conference on computer graphics, imaging and visualisation
(2006) - et al.
MatchPad: interactive glyph-based visualization for real-time sports performance analysis
Comput Graph Forum
(2012) - et al.
Knowledge-assisted ranking: a visual analytic application for sport event data
IEEE Comput Graph Appl
(2016) - J. Gudmundsson, M. Horton, Spatio-temporal analysis of team sports – a survey, CoRR; abs/1602.06994....