Machine learning that enhances the Strava User Experience and helps athletes improve their Performance. For example, we can notify you when the weather and your fitness are good for a Strava Segment PR!
Strava and Garmin Connect have a wealth of data that represents tremendous value to you as a Endurance Athlete. The achievement data we generate (Peaks and Segment Watts/kg) also serves as training data to enhance your experience through our analytics. Our mission is to unlock this rich data to help improve your user experience and help you become a better athlete.
For example, one of the algorithms, Bike Classifier, predicts which bike you rode on an activity and alerts you if your didn’t switch from the default in Strava. Another called Factors, analyzes your Peak Thresholds and Best Efforts to determine what variables are making a positive and negative impact on your results. We also have clustering algorithms, Athlete Finder, that can help you find athletes with similar power to weight ratio in your geo. Our Threshold Anomaly detection can alert you to situations where your heart-rate or power stream is inconsistent with your past results.
The Segment Performance Analytics enables three predictions for each “Watched” Segment:
a) Is the weather good for a Segment Effort? This is expressed as a z-score index with positive being good conditions and negative being bad.
b) Is your current power curve enough to bring a new PR based on the weather? This is expressed as a % attainment of your last PR speed (requires power meter).
c) What watts/kg would you have to deliver to get a new PR (requires power meter)?
For a complete list of the data science features within strive.ai visit our Roadmap page as new features being added every month.