Search

The present work introduces floodlight, an open source Python package built to support and automate team sport data analysis. It is specifically designed for the scientific analysis of spatiotemporal tracking data, event data, and game codes in disciplines such as match and performance analysis, exercise physiology, training science, and collective movement behavior analysis. It is completely provider- and sports-independent and includes a high-level interface suitable for programming beginners. The package includes routines for most aspects of the data analysis process, including dedicated data classes, file parsing functionality, public dataset APIs, pre-processing routines, common data models and several standard analysis algorithms previously used in the literature, as well as basic visualization functionality. The package is intended to make team sport data analysis more accessible to sport scientists, foster collaborations between sport and computer scientists, and strengthen the community's culture of open science and inclusion of previous works in future works., Comment: 5 pages, 1 figure. For associated package, see https://pypi.org/project/floodlight/. For source code, see https://github.com/floodlight-sports/floodlight . For documentation, see https://floodlight.readthedocs.io

The automatic detection of events in complex sports games like soccer and handball using positional or video data is of large interest in research and industry. One requirement is a fundamental understanding of underlying concepts, i.e., events that occur on the pitch. Previous work often deals only with so-called low-level events based on well-defined rules such as free kicks, free throws, or goals. High-level events, such as passes, are less frequently approached due to a lack of consistent definitions. This introduces a level of ambiguity that necessities careful validation when regarding event annotations. Yet, this validation step is usually neglected as the majority of studies adopt annotations from commercial providers on private datasets of unknown quality and focuses on soccer only. To address these issues, we present (1) a universal taxonomy that covers a wide range of low and high-level events for invasion games and is exemplarily refined to soccer and handball, and (2) release two multi-modal datasets comprising video and positional data with gold-standard annotations to foster research in fine-grained and ball-centered event spotting. Experiments on human performance demonstrate the robustness of the proposed taxonomy, and that disagreements and ambiguities in the annotation increase with the complexity of the event. An I3D model for video classification is adopted for event spotting and reveals the potential for benchmarking. Datasets are available at: https://github.com/mm4spa/eigd