About this Event
Sports Data Analytics and Visualization
Varaprasad Rao Kurra
Advisor: Dr. Ying Zhu
Sports data analysis and visualization can be a useful tool for gaining insights into games. In this project, I explored new techniques to analyze and visualize shot patterns in tennis matches. Tennis is a complicated game that involves a rich set of tactics and strategies. Current tennis data analyses are usually conducted at a high level that often fails to show patterns and nuances embedded in low-level data. This project provides new tools and techniques for micro-level analysis. The outcome of this project may be applicable to other areas too.
My project uses a machine-learning approach to understand attributes that contribute to the prediction of the serve direction in tennis. This involves obtaining meaningful patterns from tennis data that has been collected over many years. First, advanced data cleaning was done on the huge dataset. Next came feature extraction and identification of the serve direction attributes, followed by detailed exploratory data analysis and feature selection on the data. Finally, we applied machine-learning algorithms to predict the serve direction. We studied the classification of the results, using a different set of bagging, boosting, and neural-network classifiers to predict the serve direction.
Committee
Dr. Ying Zhu (chair)
Dr. Yanqing Zhang