How big data is improving results in high-performance sport

Big data & sports
Over the last decade, the analysis of an increasing volume of data has profoundly changed professional sport. It makes it possible to spot talent, to prevent injury, or to study the competition. Adrien Sedeaud, a researcher at the French National Institute of Sport, Expertise, and Performance (INSEP), explains.

What are the benefits of data in high-performance sport?

Firstly, data mining can foster talent identification. It is not a question of eliminating certain athletes, but that of not missing future talents who may slip through the net of current talent identification systems. Data analysis enables recalibration of performances according to maturity and stage of puberty, sporting history, or progression curves.

Competitive analysis is another field into which data can bring added value. How do the opposing teams win and lose?

Data analysis also makes it possible to monitor training according to volume, intensity, density, recovery method, fatigue, sleep, mood, or stress. This time, the aim is to optimize and personalize programs depending on recovery capacity but also to work on injury prevention. An “Athlete Management System” (AMS) can thus measure what an athlete actually endures compared to the initial theoretical plan, then regulate the following sessions according to the athlete’s form.

Data thus helps to better weigh up the competition

Indeed, competitive analysis is another field into which data can bring added value. How do the opposing teams win and lose? What victory patterns can be identified then modelled? Of course, other elements can be measured, objectified, modelled, and visualized through data capture or generation, such as sport motion analysis or the creation and modelling of digital phenotypes.

Does this data mining have its limits?

The field of possibilities is vast and the feedback on which to capitalize is large. However, it is necessary to be aware of the limits of what data can bring. There are limits which can be human (time-consuming activity, cognitive bias), technological (measurement and restitution biases), conceptual (prediction errors), regulatory (GDPR), or even ethical.

Among the cognitive biases, is interpretation bias in particular. Different personnel will take home different lessons from the same results. As for personal data protection, anonymization techniques have their limits. It is easy to identify who is hiding behind a certain record made on a certain date. For example if I say, “9 point 58 seconds in 2009”, you will think straight away of the athlete who achieved this performance.

Which clubs and federations are pioneers in this field?

Private professional structures such as the American leagues (NBA, NFL, NHL, MLB) are those which have the greatest maturity and the longest experience in this field, as well as the largest budgets. These American football, basketball, or hockey franchises were all pioneers.

Then follow British then European football, and rugby from English-speaking and Southern hemisphere countries. For the 2012 London Olympic Games, the United Kingdom, via its national Olympic Games organizing agency, created a dedicated centralized system for collecting and using data, with substantial staff, that is continued today.

What about the maturity of French sport in this area?

Our Latin culture can sometimes hamper the exploitation of data. Yet, the French rugby, skiing, or swimming federations are far from lacking compared to their foreign counterparts, having had data structuring for decades. They are equipped with scientific support cells that employ data scientists.

On a national level, the field has grown rapidly over the last five years. Created in 2020, the Sport Data Hub is to provide French athletes with a competitive advantage for the 2024 Paris Olympic Games and beyond. It is also about spreading the data culture and fostering the emergence of sharing and practice communities.

In anticipation of the Paris Games, a priority research project targeting ultra-high performance with a budget of 20 million euros has also been set up to accompany high-performance athletes.

One of the winning projects driven by INSEP and worth mentioning is Detect, which aims to objectify French athletes’ performances in their competition context and to estimate their probabilities of performing well at the next Olympic Games. PerfAnalytics is about identifying how video analysis can determine performance indicators, diagnose the efficiency of individual motions, and model a successful strategy of action sequences depending on context (equipment, opponents, referees, etc.).

The aim of Paraperf is to optimize wheelchair athletes’ performances with the objective of putting research at the service of paralympic athletes and their personnel to maximize the chances of a podium at the Paris Games. Finally, the Empow’her project consists in studying the impact of periods and menstrual cycles on sporting performance.

Read also on Hello Future

What’s new in PET? Federated learning


Digital identity: towards a European digital wallet?


GDPR: behind the regulation lies an opportunity for businesses


Did you know?


Innovation to create favourable conditions for digital trust


Did you know?