Strava has completed one of its largest data clean-ups yet and removed 2.3m electric bike activities from its leaderboards.
The cycling and fitness app used three machine-learning tools to address anomalous activities showing up in its results.
James, a Strava engineer, explained on Reddit that these include: enhanced ebike detection to catch activities recorded on ebikes but uploaded as regular rides; processing the top 100 activities on every global ride-segment leaderboard; and a model to better identify when a bike ride is uploaded as a run.
The result is 2.3m ebike activities have been removed, alongside 1.6m vehicle activities, and 293,000 athletes have had spots in top 10s restored.
This is the second data clean-up Strava has completed in the last year. In May 2025, Strava revealed it had removed 4.45m activities after it introduced its ‘auto-flagging’ tool Themis in February.
The tool catches activities recorded in vehicles before they appeared on KOM/QOM leaderboards by analysing 57 factors, such as speed and acceleration.
Strava said at the time that the tool had led to users flagging 72 per cent fewer activities as “in a vehicle” due to the system catching them first.
The latest crackdown fulfils Strava’s promise in February last year to investigate ebike rides on rider leaderboards.
But as James the Strava engineer wrote in his Reddit post: “Data integrity work is never really ‘done,’ and we know there’s still more to improve when it comes to removing anomalous activities.”
