Today’s competitive coaching roles are mired with seemingly incompatible skills and responsibilities. Unique sport expertise, competitive landscape, ever-evolving training methodology, sports medicine, modern technology, not to mention organizational navigation, logistics and equipment design to name a few of the key qualifications for a successful coach. Enter the data scientist and the new world of digital coaching.
When the coaching staff of USA Team Pursuit met with IBM jStart, we simply asked them, ‘how can we help?’. We didn’t pitch analytics, IoT or cognitive. We didn’t tell them about our technical prowess in these and other emerging technologies. We listened. We heard them talk about the laboratory-like environment of track racing and how they can now quantify racer improvements by a maximum of three percent annually. We listened intently as they described their lack of easy visibility into performance data and the long delays they had in accessing track metrics. We heard about the hours they put in after each practice to manually comb through stop-watch reports and power meter data. They wanted to be able to better understand in precise terms and then articulate to their athletes what a poor exchange will cost and what will be needed to make up the time. Better yet, could they provide the riders the information real time during the exchange? We got to work.
Within a month, we were back with a simple user-interface actively displaying the average pull, average non-pull and the average of both. Initially, the bar charts were enabled for a single rider; power over time. At the next training camp the first version of the mobile app carried in the back pocket of the cyclist’s skin suit was tested. The objective was to show that data could be streamed from the bike power meter and stored in the cloud. jStart enhanced the app to collect the new SmO2 data from a BSX Insight device along with the first version of an iPad app interface that displayed mobile phone and sensor device status in the collection process. Over the next few months the jStart team leveraged IBM Cloud, Watson IOT and Spark services to not only provide the coaches the analytics at the end of the session, but also in real time.
This in itself saved the coaches hours per day of manual analysis. The platform was expanded to view all riders simultaneously, thus multiplying its value. Next step, live analytics to demonstrate percentages. What is W’ (W prime)? We learned, measured and collected the data. Can you detect matches burned and provide the riders that feedback when it happens? Sure, let’s connect another device.
Perishable insight – Information needs to be delivered quickly in order to effect decision making.
This was before putting it altogether on summary visualizations to make it easy to have conversations with non-technical team members and staff. This was before we added the live-feed speed line. Wrapping it together to view it over time, the various sub-components were linked over the top of the power graph for a consolidated view of the run.
“We have a new way and a new perspective to digest and interpret data” – Coach Andy Sparks
“The biggest thing is just that constant information immediately after an effort. We are doing an effort, getting feedback, changing that effort, doing the next one. So we are not getting four efforts in, in a day, and then figuring out what we did wrong or did right that night.” – Sarah Hammer, Cyclist
“It is really important because you can immediately correct something right there in that moment instead of doing it, figuring it was wrong, coming back and trying again. It makes our time really productive.” – Ruth Windner, Cyclist
“Having the data, any sort of data and particularly biofeedback data and power data immediately available for us while we are either training or racing is extremely critical because so much of our sport involves being innovative and being able to change and react very quickly.
