Stats and storytelling: How Dimension Data is breaking down the Tour de France
CHAMOUNIX, France (CT) – In this era of smartphones, high-speed internet and ‘Big Data’, we have more information at our fingertips than ever before. This is true in many facets of life, and no less so when it comes to bike racing.
For many fans, watching the Tour de France is no longer simply about turning on the TV — it’s a multi-platform experience that involves TV coverage, a Twitter feed and a race-tracking platform.
On the eve of the 2015 Tour de France, Dimension Data and the Tour’s owner, ASO, announced a partnership that, they hoped, would increase the amount of information Tour fans had at their disposal when watching the world’s biggest race. By placing tracking devices on every bike, Dimension Data was able to collect information about each riders’ position and speed and deliver that to fans via a live-tracking web application.
A year on, at the 2016 Tour de France, the Dimension Data/ASO partnership continues. In the past 12 months, there’s been plenty of time and several other ASO races — Paris-Nice, Paris-Roubaix and the Criterium du Dauphine — for Dimension Data to expand and refine its offering. So what’s new, and what is Dimension Data’s focus in 2016 and beyond?
If you’ve been watching the Tour de France on TV this year, you would have noticed an increase in the amount of rider data being displayed on screen, and the number of ways that data is being displayed. The speeds of different riders are often shown, as too are the distances and times between the various groups on the road — useful in a mountain stage when the peloton is split right up. The distance remaining to the finish line is also being displayed in a more accurate fashion, ticking down to just a few hundred metres.
It took until stage 5 last year, but Dimension Data’s live-tracking website quickly became a useful tool for tracking groups of riders on the road during the 2015 Tour. This year, the accuracy of that tracking has been improved.
According to Peter Gray, Dimension Data’s technical team lead for the Tour de France project, this is at least partially down to the use of newer on-bike transmitters.
“They’re actually a bit smaller and lighter but they also transmit a lot further,” Gray said. “So what that means for us is that while last year we were getting a lot of good information, when we had riders that were a long way away from TV motorbikes or from any of the race cars, we’d lose that transmission.
“Now … their transmission range is more like a kilometre as opposed to 100 metres, and so it means that actually we’re getting almost full coverage of the whole peloton all day. And it means that then the tracking coverage on the website and that kind of stuff is much more reliable.”
But perhaps more important than the addition of TV graphics or changes to the online tracking software is the way Dimension Data is trying to make sense of the data that’s being captured.
While more data from the Tour de France is being shared than ever before, raw data is of little value without context. Accordingly, one of Dimension Data’s biggest goals at the 2016 Tour de France is to use the information collected to tell stories about the race.
“Having got the basics of the telemetry working, and working with increased capacity this year, we’re focusing very heavily on telling stories with the data; being able to share that information,” Peter Gray said. “How can you use that to provide better insight into the race? How can you provide a more reliable service both to the television and digital channels?”
The popular @letourdata Twitter account is back in action this year, offering interesting snippets of information about moments within the race, including:
– how the lead of breakaway groups changes over time:
— letourdata (@letourdata) July 6, 2016
– which teams are contributing to the pacemaking:
— letourdata (@letourdata) July 6, 2016
– how fast the final sprints are:
— letourdata (@letourdata) July 18, 2016
– and even how fast riders are going when they crash:
Alaphilippe fell on the descent from Grand Colombier (32km from finish) at 56km/h on a gradient of -8%. He is back on his bike.#TDFdata
— letourdata (@letourdata) July 17, 2016
Walk around the technical zone at a Tour de France stage finish and you’ll notice Dimension Data has a bigger truck at the race this year — a necessity to accommodate what is a noticeably bigger team in 2016. Among the additions is a small team of content creators that dissect race data and write daily posts with highlights from the day’s data:
Also new in 2016 is a live-tracking tool designed for commentators (but which is publicly accessible). It features many pages of useful data about the race, including:
– a text feed with updates from the race (from the main letour.fr site)
– a map with rider locations (as per the regular live-tracking site)
– bite-sized chunks of information about the current breakaway composition
– a list of all riders and their position relative to the finish and leaders (see image below)
– the various classifications, and
– a race schedule.
ASO TV commentator Matt Keenan has spent some time using Dimension Data’s commentators’ tools and while he believes the intent is right, more needs to be done before it’s a tool he’ll use consistently when live on air. As he explains, commentators are already flipping between footage of the race, start lists, their own notes and more. Spending time switching between pages of data only complicates things further.
“I’ve found it too cumbersome to navigate while I’m commentating to what’s happening on the TV monitor, looking at information about the churches, chateaus, etc and checking my own rider profile notes,” Keenan said. “The system last year, Matsport, didn’t provide as much information but it was easy to follow and would clearly show the key members of each group on the road without needing to do any navigation. I found their system easier to use.”
While Dimension Data’s commentators’ tools are only one part of the company’s offering at the Tour de France — and a part that most fans won’t see — they serve as an important reminder that having a lot of data is good, but that sometimes less is more.
“I know what they’re trying to do and I like what they’re trying to do but it needs to be simplified,” Keenan said. “The information they provide on the TV screen — riders speeds, position on the road — is the most user friendly.”
The importance of context
Some on-screen data need little explanation or further information to make sense. We all know, for example, that hitting 70km/h in a final sprint is impressive. But other data — such as the speeds of riders in different groups on a mountain stage — require a bit more context.
Consider the stage 12 Mont Ventoux climb where Chris Froome was in a small group ahead of Nairo Quintana and several others. A TV graphic showing that Froome was doing 20km/h and Quintana was doing 18km/h could lead to a rather obvious conclusion: that Froome was stronger than Quintana at that moment and was pulling away.
But what if Quintana was riding a steeper section of the climb than Froome? And what if Froome had been doing 16km/h on the section Quintana was doing 18km/h on? In order to make sense of speed data, gradient data (at a minimum) is also required.
Thankfully Dimension Data often shows speed and gradient on screen at the same time:
The collection of riders’ speed and location data allows other stories to be told as well, as Peter Gray explained.
“Some of the most interesting stuff was a few stages ago just watching Cav [Mark Cavendish] try to make the time cut,” Gray said. “We could see exactly what time gap he had to the front, which historically you’d just never be able to see that information.
“You could see it on the live tracking site and that was pretty obvious because he was in the green jersey. We highlight all the jerseys, so there’s the group with the green jersey in it and you can see what time they’re at and you can click in and have a look at Cav and see what speed he’s doing, what gradient he’s doing, feel his pain …”
Stories like these add significantly to the viewing experience of the cycling fan. And the ability to sort through rider data allows viewers to follow the stories they want to follow. The next step is to have a way of highlighting interesting stories as in the Cavendish example, and making that visible to everyone, not just those tracking Cavendish specifically.
A year on from the launch of Dimension Data’s live-tracking solution at the Tour de France the company is making good progress. More data is available than before, and that data is more accurate than it was a year ago. For Peter Gray it’s all about continual progression.
“We’re trying to provide tools that are useful for people at home, we’re trying to provide tools that are useful for the press and for the commentators,” Gray said. “[We’re] continually working to make that better and better both through the course of the Tour and also through the other races that we get involved in.”
Some of the other races Dimension Data has worked on are six-day track carnivals where the company has been able to combine the use of live on-board camera footage with additional telemetry data such as rider heartrates and cadence. While a velodrome is a much easier environment to control than the open road of the Tour de France, such developments mightn’t be far away.
“We’re already doing it,” Gray said. “At the Dauphine we managed to successfully trial that and show that it’s really just a matter of having the right riders equipped with the cameras at the right times in the environment that you want to show. That then starts to become a directorial choice, as opposed a technology limitation.”
Live on-bike cameras have been used on a couple of stages of this year’s race and the sharing of advanced telemetry data has been done in the past. But even if heartrate and power data were available to show on screen for each rider in the race, context is once again important.
Knowing a rider’s current heartrate isn’t particularly valuable if you don’t know what their maximum heartrate is. Similarly, knowing that a rider is putting out 450 watts for five minutes isn’t all that useful unless you know that rider’s weight.
For now though, it’s encouraging to see that Dimension Data has recognised that simply providing reams of data isn’t enough; that the data needs to be used to tell stories, to help shed light on the hows and whys of the sport, not just pure numbers.
Because, ultimately, that’s what sport is about — telling stories. The data is merely another way of telling those stories in an interesting and informative way and adding a level of analysis that wouldn’t otherwise be possible.