Imagine that you’re training for the San Francisco Marathon. You’d like to prepare for the distinctive pattern of hills on the course:

Now if you live in San Francisco, you could just run along the actual race course. But what if you don’t live near San Francisco? Wouldn’t it be cool if a service could suggest a running course near you that was similar to San Francisco’s marathon? That way, you might live somewhere like Portland or San Diego, but you could still prepare for the hills that you’d face in the real race.
This could be a fun/tricky problem. You could start by just dumbly comparing elevations to see how similar two runs are. A site like Strava already has GPS traces from runners, so they could compute similarity between runs. Within a selected radius, Strava could say which runs were most similar to (say) sections of the Boston Marathon.
Elevation similarity might be the easiest way to start, but Strava has other measures like “Suffer Score” that it could use instead. I’d be curious to know what running route in the Bay Area of California is most similar to the Boston Marathon, for example.
It would also be super interesting if Weather was integrated as a part of this problem.
For example, In Melbourne we see 4 different types of weather depending on which area you are in. It would be a waste to just practice in a hot area when the marathon would be in a chilly region.
It would also be interesting to gather the same information across the globe. Here in Ireland for example, it would be great to see if there’s a route that we could run in parallel and compare times… Globally!
Is it working at lake Balaton too?
It would also be interesting to gather the same information across the globe. Here in Ireland for example, it would be great to see if there’s a route that we could run in parallel and were goods
Matt, I’ve never seen a tool like this, but certainly the Strava folks would have enough data to cook something similar to this up.
The story reminds me of the 2008 Olympics Gold Medalist time trial cycling champion Kristen Armstrong. When the course route was announced, she flew to Bejing a year before the Olympics, rode the course with a GPS bike computer, then took the elevation and distance stats and eventually found (via more trial and error than programming skills in a GIS dataset) a loop in Idaho with a similar distance and elevation profile. So she rode that course every week for a year, practicing where to go faster and when to rest, week after week. She was the most prepared and crushed her competitors, being the only one to finish below 35min and almost 30sec ahead of her nearest rival (finishes separated by fractions of a second are more normal in this discipline of cycling).
I had no idea that Kristen Armstrong did that–wow! That’s incredibly cool.
The neat thing is that Strava could get some free buzz with very little work. Take every ~26 mile GPS trace west of the Mississippi with the word “marathon” in it, then just compute a similarity/hash and tell us “which West Coast marathon is most similar to Boston.” A lot of people would be interested in that.
I just ran my first(and probably last) half marathon this month in Nashville. I would be interested to see my “suffer score”!
Interesting article Matt 😀