Machine learning enters the Tour de France

Machine learning enters the Tour

Machine learning is coming to the fore across a range of professional sports. It’s helping to predict everything from injuries to peak performance levels. And, now, it’s entered the Tour de France. The world’s biggest cycling event takes 198 riders along 3,540 kilometers of the French landscape. And this year, the race is utilizing artificial intelligence (AI) for the first time ever.

How machine learning will play a role

Organizers of the 104th edition of the race have teamed up with a company called Dimension Data to introduce machine learning into the event. It’s a pilot program that aims to predict the likelihood of various race scenarios. The data could, for example, determine whether the peloton (the main pack of riders) will catch the breakaway riders and even suggest likely points in the race for this to occur.

Scott Gibson, group executive of Dimension Data marks the occasion as inevitable. “As more technology is introduced into professional cycling, the viewing experience is transforming and its popularity increases. What’s especially exciting for us is how we’re helping the organizers attract a new generation of digitally savvy fans, and how advanced technologies like machine learning are opening up new possibilities for providing the insights that today’s fans demand.”

Using GPS transponders, installed under the saddles of each bike, Dimension Data will collect a staggering 3 billion data points throughout the race’s 21 stages. That’s a significant increase from last year’s 128 million data points. The data collected from these transponders will be combined with other impacting data such as the course gradient and weather conditions to bring viewers a range of breaking stats, including live speed, the location of individual riders, the distance between riders and the composition of different groups within the race. The result is a far more immersive experience from a viewer’s perspective.

And all of this is managed via a cloud-based data center which relays these stats to broadcasters, allowing them to tell viewers even more about their favorite teams or riders. This cloud-based structure gives Dimension Data the flexibility to control the system from anywhere in the world. And they’re doing just that as the system will be operated from across four separate continents.

What about security?

As with any digitally-based system, the threat of cyber attack is real. But Dimension Data has done an incredible job in strengthening the defenses of the system. It’s likely that they learned a thing or two from last year’s Tour de France where they blocked just over 1.4 million suspicious attempts to gain access to the race’s digital systems.

Race director, Christian Prudhomme is fully behind the use of this technology to ensure that the sport keeps evolving alongside its viewers. “Today our followers want to be immersed in the event. They’re more digitally engaged and active on social media than ever before. They want a live and compelling second-screen experience during the Tour. This technology is helping us to achieve this and will completely transform their experience of the race.”

For anyone who doubts the growth of the sport, Dimension Data also released a few key figures pertaining to the race. Some 198 riders make up 22 teams. That’s a lot of data and means there’s a lot of interest. The race’s live tracking website has been boosted from supporting 2,000-page requests per second in 2016 to a staggering 25,000-page request per second this year. TV broadcasting hours are also up significantly from 80 per channel in 2016 to 105 per channel this year. All this means that the Tour is likely to smash its 6,100 hours of TV coverage it received last year.

If you’re interested to see how machine learning is infiltrating other industries, read our latest article. It outlines six other sectors that are likely to be disrupted by machine learning in the near future.