One of the initial visions for strive.ai was to offer features that help athletes find training partners with similar characteristics.  We tried one approach in the early beta that clustered athletes based on starred segment completion time.  This approach had “mixed results” so we went back to the drawing board.

We landed on using watts/kg threshold values over [1, 5 and 20] minutes as the source data for this feature.  The only downside is we lost the ability to cluster with athletes not using strive.ai.  Now that we have several hundred users this downside is less of an issue.

The initial release clusters athletes using bike power and filters out anyone that doesn’t have weight configured in Strava.  We’ll get the run power one going next, followed by best effort pace clustering for runners.

You can access this feature off the main navigation bar on the site using the Athletes option.

Athlete Cluster

The menu option can be selected at any time and will cluster each time you access it.  The images and links point to the strava.com profiles of each athlete in your cluster.  We’ll add a filter to remove people you’re already following on strava soon.

One caveat is we’ve not yet included geography in the process.  This means that you’ll be clustered with athletes from other cities.  Once we get over 1K athletes we’ll sort that out to allow you to choose across the system or only in your geo.

Now that we’ve got this base feature we’ll start brainstorming other useful things to help you find other athletes, segments, rides and races that might suit you!

Please let us know if you see any odd results or errors to mike.wood@strive.ai

Athlete Clusters