Modeling athletic performance using mathematical data science
December 06, 2021
Byrne Scholar and Mathematical Data Science major Joe Gyorda ’22, second from left, interned this summer at HALE Sports, which seeks to optimize the health and performance of athletes through analyzing somatic and biometric data. “The first team was building a computer vision model in which athletes could receive real-time feedback on various exercises (e.g., squat, pushup) to correct their form. Machine learning and body pose estimation models were implemented to analyze body position/angles and identify correct/incorrect form,” says Joe, himself a triathlete. In his second team Joe worked with longitudinal somatic survey data collected from college athletes, with the goal being to implement clustering algorithms to track athlete variables over time and using predictive models to determine which variables (e.g., energy, mood) are impacted by changes in another (e.g., sleep). “I am very grateful for the generosity of the Byrne Fund for supporting me and my work!”