Enhancing Sports Analytics through Web-Based Application: A Workflow Perspective

The rapid advancement of the integration between sport science and computer technology has given rise to a new web-based application with focus on sports analysis. The design and development process of this web application is presented in this paper, focusing on the workflow and processes involved using Python Flask framework, video processing, video manifest structure, database structure with MongoDB, machine learning back-end with Keypoint RCNN, and the front-end intuitive data visualisation interface. Insights of athletes performance are gained through the integration of frame accurate video play-back, human pose estimation, concise biomechanical reporting, and consistency measurements. Despite its primary emphasis on archery, the fundamental workings can be applied to a broader sporting implementations. Preliminary field tests with a sports centre indicate a positive reception from users, noting its efficiency and depth of analysis. This paper highlights the capacity of integrating computational tools with sport science to provide objective insights to optimise training and performance evaluation.