Data takes to the road - the technology behind the Tour de France

The world's greatest cycle race is transforming its relationship with fans thanks to internet of things, data analytics and machine learning technologies

This article can also be found in the Premium Editorial Download: Computer Weekly: Tech takes the lead in the Tour de France

As recently as 2014, the only way to find out real-time information during the Tour de France was from a chalkboard, held up by a race executive sitting as a passenger on a motorbike driving ahead of the cyclists.

For fans accustomed to enhancing their enjoyment and understanding from a deluge of data when watching football, cricket, tennis and other high-profile sports, the world’s greatest cycle race was something of a challenge to watch.

TV viewers could see the timings, and watch numerous camera angles, but the subtleties and tactics of elite professional road racing were a mystery to all but serious cycling enthusiasts.

That’s all changed. At this year’s Tour de France, fans have access to more data than ever before, explaining and analysing the performance of Britain’s Chris Froome or Mark Cavendish in intense detail across three weeks of racing.

But unlike those other sports, which are typically contained in a single venue, “Le Tour” presents a unique set of challenges for the technology that now relays data from bike to TV viewer in two seconds.

The tech behind the Tour is provided and managed by IT services firm Dimension Data – which separately also sponsors the racing team that Cavendish rides for, the first African outfit to compete in the Tour, now known as Team Dimension Data.

Aside from the usual promotional benefits of a sponsorship association with one of the most popular sporting events in the world, for the supplier, the conditions it has to tackle during the Tour provide a learning experience for how to handle high-volume, real-time data from internet of things (IoT) devices in often harsh conditions.

We need to deploy strong and robust technology – yesterday we were in the mountains, today we are in another town. Other sports are making more use of data and we need to do the same
Pascal Queirel, CIO, Amaury Sports Organisation

Each of the 198 riders that start the race has a bespoke sensor attached to a clip below the saddle of their bike. The sensor, which weighs just 100g for a weight-obsessed sport, contains a global positioning system (GPS) chip, a radio frequency (RF) chip, and a rechargeable battery with enough power to last the longest of the Tour’s 21 stages.

Each sensor transmits its GPS positioning data every second – producing over three billion data points during the course of the race. Dimension Data also takes feeds from other sources such as weather services and road gradients, and combines the real-time rider information with historic data gathered since the firm became the tech provider at the 2015 Tour.

As a result, race organisers, teams, broadcasters, commentators, TV viewers and fans using the Tour de France mobile app, now have access to in-depth statistics on progress of the race and their favourite riders. The riders wear earpiece radios so their teams can relay real-time data to them while they cycle – no need to keep an eye out for that chalkboard anymore.

Race insights

Even diehard fans have been surprised by the insights gained. When Dimension Data released a graphic (below) showing the speed and sudden deceleration of the bikes involved in a high-speed crash during the 2015 Tour, sceptical fans pointed out what seemed to be an obvious error – the data suggested some riders had somehow accelerated after the moment of impact.

Further analysis showed that the bikes themselves – with the sensor still attached – were briefly catapulted forwards as the rider was thrown off, such was their momentum. Some bikes that seemingly accelerated to race speeds soon after were, in fact, picked up by team cars following the peloton of cyclists and driven away.

Similarly, when data suggested that some riders were achieving top speeds above 90km/h (56 miles per hour), some fans were doubtful, until Cavendish tweeted a photo of his speedometer showing he peaked at over 100km/h (62 miles per hour).

There can, however, be times the technology gets it wrong. GPS data can be corrupted by signal noise during transmission. As a result, during the 2015 race, one cyclist was miraculously transported from France to Kenya, according to the data. Today, all data is checked and cleaned for such errors before processing.

So while even diehard fans have found fresh insights from the data analytics, the technology challenges to get it to them remain.

Technology challenges

Tour cyclists traverse some of the most difficult terrain in sport – up and down remote mountain roads in the Pyrenees and the Alps, where mobile phone signals can be rare and weak.

As a result, data feeds from the bikes use the white space in television signals to transmit back to the Dimension Data technology team. GPS data is transmitted using a mesh network via antenna on race cars following the cyclists, up to a swarm of TV helicopters overhead. From there, signals are relayed to an aircraft further above, and then sent to the TV trucks at the race finish. There, the data is split back out from the TV feed and picked up by Dimension Data’s “big data truck”, which follows the race around the French countryside every day, waiting at each stage finish.

The transmission distances and latency of using satellites is too slow for the Tour, says Dimension Data chief technology officer Ettienne Reinecke. He says the firm is testing the use of low power, wide area wireless networking, based on Cisco’s LoRa technology, to create a “wireless bubble” around the riders – but that’s for a future race.

Read more about technology in sport

Data collected during each stage is collected in Dimension Data’s cloud service, based in datacentres in London and Amsterdam, where data analytics algorithms combine with external feeds to produce the real-time information sent out to broadcasters, social media and the race app. The whole process from bike to viewer takes just two seconds.

The tech setup would not have been possible without the emergence of the cloud, given the remote nature of much of the race and the fact it changes location each day.

“The difficulty is we are moving every day,” says Pascal Queirel, CIO for Amaury Sports Organisation, the organiser of the Tour de France.

“We need to deploy strong and robust technology – yesterday we were in the mountains, today we are in another town. Other sports are making more use of data and we need to do the same.”

Digitally engaged fans

The technology is opening up the Tour to new and old fans like never before – for example, in 2014 video clips put out by organisers had six million views; in 2016 that grew to 55 million.

Christian Prudhomme, director of the Tour de France, says: “Today, our followers want to be immersed in the event. They’re more digitally engaged on social media than ever before, and want a live and compelling second-screen experience during the Tour. Technology enables us to completely transform their experience of the race.”

Technology enables us to completely transform fans' experience of the race
Christian Prudhomme, director of the Tour de France

This year, Dimension Data introduced machine learning capabilities for the first time, says the supplier’s Tour de France technology leader, Peter Gray. Using algorithms devised by a team of cycling experts, data scientists and sports analysts, the system makes predictions before and during the race, based on current and historic data on riders and the state of the race.

“Machine learning is the true ‘brain’ of our computing capability. It’s how our advanced data analytics platform correlates and integrates different sets and types of data,” he says.

“Through the complex algorithms we’ve programmed into the platform, the solution can then calculate the likelihood of possible race events in real-time, such as whether the peloton might catch up to breakaway riders or not, as well as create rider profiles based on their performances in past races.”

After the first 12 stages of this year’s Tour, Gray says the system correctly predicted the stage winner six times.

Unlike other sports, where the action is typically contained in a single day or event, the tech behind the Tour is being constantly updated during the race. Dimension Data has a DevOps team to allow a 24-hour development cycle, with overnight enhancements such as new data visualisations written in Australia and tested at its Hong Kong cloud datacentre, before being sent to production in London and Amsterdam ready for the next day’s racing.

Cultural challenges

Further use of data is likely in future – but technical and cultural challenges have to be overcome first.

Test have been carried out with riders wearing GoPro cameras to feed live race footage, potentially offering fans a cyclist’s eye view of the action. But, according to Gray, the difficulty of battery life, weight, connectivity and bandwidth availability during the race has yet to be overcome.

Riders wear biometric measurement technology – advanced forms of Fitbit-type devices – during the race, to monitor their physical performance. But that data is currently kept private by the teams – imagine the competitive advantage to another top rider if they knew their rival wearing the leader’s yellow jersey was showing a heart rate well above normal, suggesting they are near their physical limits.

Team Dimension Data trialled publication of biometric data such as heart rate, power output and cadence, during the 2016 Tour of California, but currently there is no agreement among teams to make such information widely available.

The Tour de France was first run in 1903, established as a marketing exercise to increase sales of a French sports newspaper, L’Auto. Over a century later, the world-famous cycle race is hoping technology and data analytics will help to engage with a new generation of digital fans.

Read more on Internet of Things (IoT)