The 2016 Olympics in Rio was labelled the most technologically advanced games to date. How are advances in gathering and analysing data being used to improve athlete performance?
LexisNexis recently published the third podcast in the Small Data Forum series, where I discussed the growing role of data analytics in professional sports with Neville Hobson, Senior Business Consultant at IBM and Sam Knowles, Founder & Managing Director of Insight Agents.
The 2016 Olympic Games in Rio saw Team GB placed second in the medal table – behind the US and ahead of China – with the cycling team winning six golds and four silvers. Competitors from Australia, France and Germany hinted that Team GB cycling team may have used unfair competitive advantage to achieve such a result.
Team GB cycling is well-known for embracing the philosophy of ‘marginal gains’: analysing anything and everything to create a competitive advantage. This includes everything from the made to measure clothing the athletes wear to the bespoke bikes they ride.
Recent developments in data collection, data analytics and machine learning have made it possible for this process to be analysed in near real time. Subsequent action, such as adjustments of an athlete’s setup, can then be taken in near real time to improve performance.
Marginal gains
There has always been data in sport. Two thousand years ago competitor rankings allowed Romans to place bets on gladiators based on previous performance. The modern data landscape is far more complex, but it can also offer far greater insight into performance.
More data points are available and how they are harnessed to improve performance is an ongoing process: marginal gains; systematic data gathering; physical tests; technology and materials analysis are some of the techniques that are beginning to demonstrate the importance of data in professional sport.
Small insight from big data
The focussed application of technology in professional sport means more data can be gathered to provide insight into small, incremental steps that will improve performance. Central to this concept is data analysis – distilling large data sets down into smaller chunks that generate actionable insight.
In the episode #3 of the Small Data Forum podcast, Neville Hobson points to the example of the US Olympic cycling team working with IBM’s Watson – a technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data.
Watson provided the US cycling team with data analysis that enabled real-time strategic insight. Historically, this was a manual process and it would take days before insight could be gained. Statistics needed to be analysed offline and the results interpreted.
Watson gathered data from cyclists and analysed it as they trained, through a combination of sensors, a smartphone to transmit data from the athletes to a computer and a tablet computer to view analysis as the cyclists rode around the track.
Hundreds of metrics were tracked and analysed, including vital signs, track position and the amount of sleep a rider had got. The analysis even considered the volume of hair on an athlete’s legs.
All of this seemingly unconnected data is then transmitted from smartphones for real-time analysis by Watson. Adjustments were made based on the data presented in a visual dashboard.
Predictions for the future
The US Olympic cycling team was the first to experience the effect of real-time data insight during an Olympic Games, however it will not be the last. While technology is essential in this process, human intervention is still required to interpret results and decide which metrics are important, where the focus should be and which metrics should be ignored.
It is the availability of such a wide range of data, the speed of collection and analysis as well as the ability to visualise data – in a form that lets the decision maker make critical decisions about performance and strategy – that is making insight actionable.
The role of coaches is now to glean insight from already analysed data and apply it to individual and team performance.
It is still human athletes, though, that other humans come to see perform at Olympic Games and any other sports event. Big and small data analytics help optimise the conditions, continuously improving by small increments. On the day it’s the athlete who wins the medal, however.
Not a robot, or an algorithm.