Data science and sports: a winning combination
Athletes always strive for the top. How can data scientists assist them in improving their performance? During the seminar Data Science and Sports, the possibilities and challenges of collaboration between these two worlds were discussed.
An interesting mix of speakers gathered at the TU Delft last Thursday, 7 April. Among them were former Olympic athlete Kamiel Maase (NOC*NSF), speed skating coach Jac Orie (LottoNL-Jumbo) and many researchers and data scientists from Leiden University, TU Delft, Qualogy, Gracenote and other organisations.
‘Because we want to win’
During the seminar, several fruitful collaborations between sports and data science were mentioned. In different sports (football, sailing, rowing and speed skating), many different kinds of data are gathered - from wind and tides to tactics, training programmes and health data. The goal, however, is always the same: analysing the data in order to discover patterns, which will help improve athletes’ performances. ‘Why do we do it? Because we want to win’, as Jac Orie simply put it.
Training
Koen Muilwijk (InnoSportLab), for instance, explained how data science is used in sailing by applying near realtime data collection and visualisation. LIACS researcher Arno Knobbe showed how he assists speed skating coach Jac Orie (LottoNL-Jumbo) in optimizing training programmes for his skaters. And TU Delft student Carli Wensveen gave a presentation on her graduation research, in which data science is used to determine different football teams’ tactics.
Garbage in, garbage out
Moreover, some interesting tools for predicting sports outcomes were discussed (Gracenote, Qualogy), as well as the latest innovations and applications in the field of Virtual Sports (TU Delft).
But there was also room to address the challenges of data science in sports. Kamiel Maase (NOC*NSF) mentioned a few, such as privacy issues, data stewardship challenges (‘garbage in, garbage out’) and the mixed nature of sports data sets.
Sport Data Center
Collaboration, however, may be the key to facing these challenges. In the final presentation of the day, Joost Kok (Leiden Centre of Data Science) introduced the Dutch Sport Data Center (SDC), in which TU Delft, LUMC, Leiden University and AISS jointly work on the application of data science in sports. The SDC has several different research lines, such as Elite Sports, Adaptive Sports & Rehabilitation, and Fraud & Risks.
(JvdB)
The seminar Data Science and Sports was organized jointly by Delft Data Science, the Leiden Centre of Data Science (LCDS) and the Sports Engineering Institute of TU Delft, and took place on 7 April, 2016. All presentations of the seminar can be downloaded here.