👋 I’m a solo builder & runner from Seoul. I built this tool because Strava’s numbers alone couldn’t tell me if my training was actually improving or not.
Hi there! I’m an IT professional based in Seoul who also loves running.
After running for about 3 years, I had a thought: Since I have all these records on Garmin, couldn't I use them to understand what kind of runner I really am?
So, I decided to dig into the data myself. I downloaded my GPX files from STRAVA(or any platform like Garmin that exports GPX)and started calculating the numbers. I analyzed how much my pace varies during a race and how much elevation changes (uphill/downhill) affect my performance.
I got a bit ambitious and integrated a weather API as well. I’ve always felt that I’m quite sensitive to the weather, and I wanted to see if the data backed that up.
I used AI to interpret the analyzed data. (I was using gemini-2-flash-lite, but the API costs were getting too high, so I switched to the gpt-oss-20b:free model.)
The AI gave me some meaningful insights. For example:
So, I decided to take this tool to the next level. Analyzing just a single record felt insufficient, so I expanded it. I’m currently testing three different analysis modes:
I also asked the AI to propose training methods based on these analyses. Anyway... I’m constantly tweaking and testing things.
Lately, I’ve been completely hooked on this side project and am steadily upgrading it. (To be honest, my dream is for this to turn out well enough to hand over to Strava someday. Haha.)
To make that happen, feedback from actual runners is crucial. Right now, I’ve built an automation workflow using Tally (for data entry) and n8n. If you enter your data, the system analyzes it and emails you a report.
If you are a runner, please give it a try and share your thoughts. Any feedback is welcome. Right now, I only have opinions from myself and my friends, so the perspective is limited. Even negative feedback (constructive criticism) would be incredibly helpful for improvements!