London Marathon: Improve your training with the IoT

By and Rich Howarth | 5 minute read | April 12, 2017

Marathon runners

To many, the arrival of April means better weather, Easter and perhaps a half-hearted attempt at the couch-to-5k app. To serious runners up and down the UK, however, it means the London Marathon is upon us at last. So this is one for you runners, you brave and wondrous folk.

We asked IBM-er Rich Howarth, a frequent marathon runner, for his ideas on how the IoT could improve training for anyone attempting the London Marathon at the end of the month. Besides being a runner himself, Rich is also VP of Engineering on the Watson Data Platform, so he’s ideally placed to offer some expert insight. Take it away, Rich.

Heartache, injury and the limitations of training books

I’ve run multiple marathons over the years including the New York Marathon on November 6th with 51,387 other people, most of whom were running this amazing event for fun. Like other runners, I have read many books on marathon training programs. Most of these are written by gifted professional runners, based on their experiences as runners and coaches, and geared toward achieving a personal-best finish time.

The catch is, they are not based on data collected from average runners, nor geared toward what most average runners want: a training schedule they can follow that leads to an enjoyable marathon without injuries.

Two of my IBM colleagues also ran New York, and unfortunately ended up struggling with injuries. This is a common and heart-breaking fate for too many recreational runners. We enter lotteries to get coveted race entries, devote months to training, and cover hundreds of miles, only to wind up injured, in terrible pain, or even unable to finish the race. Being a data and technology nerd at heart, this made me wonder how IoT and data analytics could help.

To make a real start, we need to capture data in real time

Anyone running a marathon is probably going to have a GPS watch and/or phone that can measure splits (minutes per mile or km), heart-rate, and maybe even advanced things like cadence (steps per minute) and blood-pressure. High-tech runners may have sensors in their clothing or shoes. The 2016 NYC Marathon tested biometric sensors on the shirts of 10 athletes, which captured real time data on heart rate, running cadence, breathing rate and more. This is a start, but we can do so much more!

If even basic data could be captured for a large portion of runners during the race, and the preceding 3-6 months of training, and supplemented with height, weight, experience level and any injuries, it could be used by many purposes. It would help us understand what type of training programs and distances produce the best results for different kinds of runners.

Use case: Team USA and Watson IoT

This incorporation of real-time data analytics from the IoT into an athlete’s training isn’t just pie in the sky – it’s already happening. In preparation for last year’s Olympics, USA Cycling Women’s Team Pursuit joined forces with Watson IoT, IBM Analytics and IBM jStart to capture individual performance metrics for each cyclist. The data allowed the riders to train smarter, recover better, and have more confidence in split second decisions.

With the insights from data captured during training, Team Pursuit shaved more than 2 full seconds off their results and went on to win Silver at the 2016 Olympic games.

How far can running apps take us?

As yet, a sophisticated tool like the one used by Team Pursuit isn’t available for solo marathon runners, though running apps offer a good starting point. Today, data captured by Runkeeper, Mapmyrun and other apps can show high-level correlations such as miles per week and predicted marathon finish times. However, more detailed information required to minimize injuries is not generally available.

For example, marathon coaches debate the ideal length and frequency of the longest runs—should they be limited to 16 miles or is something above 20 miles beneficial? Since a 20-mile run takes longer than three hours for most people, some experts believe this leads to injuries and should be avoided, while others feel it is critical for preparing your body for a marathon. I haven’t seen any data supporting either hypothesis.

We need an IoT and analytics solution

An IoT and analytics solution, especially one employing cognitive capabilities, could collect and aggregate training and race information (ensuring it is all anonymous) and deliver incredible insights into ideal training regimens that yield successful, injury-free marathons. These insights could be used to build customized training programs that improve the whole marathon experience, from first training run to race-day finish line. We could even a build predictive models to help people understand how prepared they are and what to expect on the big day, whether they are running in the London Marathon, following Pheidippides’ footsteps in Athens, or taking part in other events around the world.

IoT for runners with disabilities

Marathons are hard work even if you’re physically fit. But what about runners with disabilities? Can the IoT offer practical help? Last year, IBM Bluemix worked with blind ultramarathon runner Simon Wheatcroft to develop an app that would guide him safely through an ultramarathon course. The eAscot app (named after Simon’s guide dog) uses a combination of sensors and satellite navigation to plot an optimal route, and alert Simon if he veers off course through a series of short beeps.

What if you could configure this app with venue-specific data (the London Marathon route map, for instance), and help keep a blind or partially-sighted runner on course, without the need for a human guide to run alongside them?

You could even have an audio map that would alert runners aurally to the position of the next drinking station, or allow their friends and supporters to temporarily share their location with the runner’s map app so that they’d know a friendly face was just around the corner.

Learn more and share your thoughts with us

If you’d like to learn more about the eAscot app for blind marathon runners, you can read Simon’s story on our blog. To learn more about our work with USA Cycling Women’s Team Pursuit, check out our full coverage of the collaboration.

If you are a marathoner, data nerd, IoT developer, running technology supplier, or race director and want to join a project to make this happen, comment below. Maybe we’re onto something.