Using Big Data to Boost Athletic Performance

Sports analytics is a rapidly developing field that harnesses large amounts of information to make data-driven decisions about athletes, strategy and organizational management.

It is fundamentally changing the way sports are managed, played and consumed. As sports analytics continues to evolve, the demand for new models that can more fully analyze athletes' performance and training data is growing.

David Clarke, a Professor in Simon Fraser University's Department of Biomedical Physiology and Kinesiology, directs the Laboratory for Quantitative Exercise Biology. His lab develops models that predict the body's adaptations to exercise training, which are then used to optimize training programs for improving health and athletic performance. 

Much of sports analytics research to date focuses on developing empirical models that relate athletes' performance data to publicly available results and athlete statistical data. Clarke and his team apply a complementary approach by creating models that consider physiological variables. Such models are beneficial because they can pinpoint the factors that most limit an athlete's performance. This knowledge can help tailor the athlete’s training program.

To create these models, Clarke and his team collect and process massive amounts of data. A challenge they faced is that regular desktop computers provide only a fraction of the computing power needed to easily analyze and model these data. Clarke therefore worked with SFU's Big Data Hub to access high-performance computing infrastructure via Compute Canada in order to handle their computing needs as well as provide technical expertise that helped move their project forward.

"The Next Big Question fund and research support from SFU's Big Data Hub have been essential to our ability to conduct this research," Clarke says.

Clarke's project is partially funded by the Big Data Hub's Next Big Question (NBQ) Fund, which encourages researchers from across all disciplines to put forward research questions that expand our understanding of big data and transform how we use data to accelerate discoveries and mobilize knowledge. The NBQ Fund supports projects like Clarke's to empower SFU researchers to apply big data methods to their work that positively impacts communities, industries and government.

This project represents the first substantial collaborative project between SFU and the Canadian Sport Institute Pacific (CSIP), a non-profit that provides comprehensive support for Olympic and National sports to enhance athlete performance through services like sport science, sport medicine, administration and business operations. This collaboration will help CSIP further develop their sports analytics capabilities and provide SFU researchers and students with rewarding research and career opportunities.

Clarke hopes that this project, and his lab in general, can continue to innovate in the rapidly evolving field of sports analytics. "While many teams collect vast amounts of data on their athletes, methods for fully exploiting these data for performance and training insights are lacking," Clarke says. "Our work is dedicated to developing new models that can be incorporated into software that coaches and athletes can use to gain these insights. The benefits that we have realized with this partnership underscore SFU’s mission to be Canada’s engaged university and help validate the university’s strategic decision to focus on big data as an important conduit for doing so."