A new dataset for spatial, temporal, and tactical analysis of female handball player movements

Abstract

This paper presents a comprehensive dataset detailing the precise indoor positioning of players from a female amateur handball team across 10 real matches. Utilizing Ultra-Wideband (UWB) technology, the dataset captures each player's x and y coordinates every second throughout the games. Additionally, a preliminary game analysis is included, specifying the initiation and termination times of each team's possession. This analysis provides a variety of labels crucial for training machine learning algorithms, encompassing distinctions such as attack or defense, structured or unstructured play, goal outcomes, and the differentiation between counter-attacks and static phases. In total, the dataset comprises 84691 positioning measurements, offering a valuable resource for in-depth study and analysis of player dynamics and game strategies in female handball.