There is a lot that goes into the calculation.
Calculating exactly what kind of ingredients and how many calories each cyclist needs is no easy task. In the past, team nutritionists would start preparing their calorie estimates weeks or even a month before the start of the Tour de France. To begin with, they would look at the stage profiles and the weight and body composition of the riders. But when the race begins, there will always be unforeseen factors. The weather, changing team tactics and many other things would require a quick reassessment of the entries.
Asker Jeukendrup, Jumbo-Visma’s head of nutrition, said in interviews that he would have to create spreadsheets for each rider to calculate their optimal intake and turn it into meals.
🇫🇷 #PTO2022
Pre-race talk 🗣
What are they @rogla Y @LAPORTEChristop talking about? 🤔 pic.twitter.com/EVS3DtzvWX— Jumbo-Visma team cycling (@JumboVismaRoad) July 12, 2022
“It was time consuming and I could only do the calculations for 2 riders at a time. Slowly we scaled up, but when the Jumbo Foodcoach app started doing the math for us and translating it into meals, that’s when it got scalable and faster. Jumbo has been a great partner in this and we continue to improve the app.”
collecting data
Improving and creating an app that would help automate the process was a big step forward for the team. For this to work effectively, they needed to make sure the app had all the data it needed to create accurate predictions. This is what they included.
- A Garmin device on each cyclist’s bike that shows real route data on total distance, meters climbed, etc.
- A crank based power meter that provides an accurate calculation of calories burned.
- Weight, height and individual role of the cyclists (sprinter, climber, etc.).
- Weather forecast in combination with GPS location data for each cyclist to help calculate the impact of weather (tailwind or headwind, etc.).
The collection and visualization of this data is powered by Smartbase, a data management and analytics platform. Jumbo-Visma trainers would use this platform to input the actual data.
supervised learning
The team also needs to clean up this data by removing errors. For example, if a cyclist forgets to turn off his Garmin device at the end of a stage, he should exclude the post-stage portion of the data. They also relate certain variables, such as horsepower, energy and elevation, to make it easy to compare stages and races. All of this data is then used to prepare forecasts.
Using training examples, Jumbo-Visma applied supervised learning to teach its algorithm what the output of calorie predictions should be. This was a regression problem, and they selected random forecasting as the best machine learning algorithm to solve it.
Machine learning is much more accurate
This all sounds great, but it doesn’t count for much unless the predictions are at least as accurate as when they are made manually by human nutritionists. They used R-squared to test both their machine learning model and their manual predictions. R-squared measures the strength of the relationship between the model and calories on a 0-100% scale. The machine learning model scored 82%, while the manual predictions only scored 52%. This is really amazing! Not only is the algorithm more accurate, but nutritionists also get results in a fraction of a second, giving them more time to respond to unforeseen factors that arise during races.
The Jumbo Foodcoach app translates it into meals
The Jumbo Foodcoach App connect the dots. When a team nutritionist enters the calorie numbers from the machine learning model into the app, it provides sample meals with optimized ratios for each meal. It can’t get much easier than that. That’s one of the big reasons the team keeps winning. You may check out their app in the App Store or Google Play
