I may cross-post this to my general-purpose blog, because it’s just geeky enough to warrant a deeper explanation of the details.
As the nature of my work changes and [hopefully] morphs into what will ultimately be my PhD thesis, I spend a lot of time tinkering with new methods. Without delving into any appreciable details, I’ve been experimenting with a new one called Gaussian Processes. And what better way to experiment with them than on my very own running data!
I downloaded all my data (233 runs!) from Garmin Connect and wrote a script to track my progress by computing an average pace for each run. Then it does some cool regression to determine any patterns underlying the behavior, and it spat out this graph:
Pretty cool! There are tons of simplifying assumptions, but this seems to indicate that my average pace is getting faster over time. Not by a lot; you can definitely see a nice spread of average paces all over the place. However, it’s worthwhile to keep in mind that this mixes together races, tempo runs, speed workouts, long runs, and short easy runs, all of which have their own characteristic pacings, so it’s not like I’m going all out every time I hit “start” on my Garmin watch. Still, it’s cool to see an overall progression towards a better pace!
Yeah that’s really all I wanted to say. I’ll post a more in-depth discussion on my other blog soon 🙂