Women’s Soccer Penalty Kick Data Captured by Computer Vision Models

The convergence of sport and data science into sports analytics provides granular data useful for strategy and research. A common research area is the soccer penalty kick. While typically categorized by outcome, penalties contain detailed biomechanical information in the lead-up to the kick that is often ignored. Modern tracking technologies can now capture these complex actions. In this dataset, YOLOv8 computer vision models were applied to video clips of 132 penalty kicks taken by former collegiate women’s soccer players. The first model recorded penalty takers’ poses across 12 keypoint markers (shoulders, elbows, wrists, hips, knees, ankles) at roughly 25 frames per second, while the second model identified the contact frame and kick direction relative to the goal's center. In addition to demonstrating the utility and efficiency of computer vision systems, the resulting dataset offers new avenues for exploring penalty kick strategy, kicking mechanics, and game theory in a rare women’s sports setting.