Road Gear

How to predict Tour de France winners with data

Dimension Data has added a new predictive feature to its Tour de France coverage that may help you pick a winner.

Want to know who will win the next stage at the Tour de France? Dimension Data has a new real-time predictive data feature to its streaming Tour de France coverage. It provides the odds and possibilities for different race outcomes.

The information and technology services company that sponsors Mark Cavendish and the Dimension Data WorldTour team first partnered with the ASO for the 2015 Tour de France. It provides real-time rider data like speed, power, and GPS locations on the course. Small GPS devices installed on each rider’s bike capture key information. This is processed and analyzed in real time for cycling fans, commentators, and TV broadcasters during the race.

New this year, Dimension Data digs even deeper into the numbers with machine-learning and predictive analytics. Using complex algorithms, the company says it can calculate the likelihood of possible race outcomes in real-time. Will the peloton catch the break? Or will the small group out in front stay clear to the end? Dimension Data will calculate the likelihood of each outcome.

Photo: Dimension Data

The new predictive analytics use the same data collected and distributed from the on-bike GPS trackers. This is combined with third-party environmental data like the road gradient and weather conditions. It then integrates historical data collected by Dimension Data’s live tracking over the past two years as well as rider performances, stage profiles, and race statistics from the last five years.

Dimension Data will provide data-generated rider profiles that show each rider’s strengths and weaknesses against the stage profile. It will also provide the statistics on possible finishes.

Real-time data and predictions are available on Dimension Data’s live-tracking app. You can access heat maps and speed comparisons for each stage of the Tour.

Read more about Dimension Data’s predictive race data here >>