Jekyll2025-04-22T02:22:18+00:00/feed.xml3thirty3:30 - Almost there.Ethan SmithProductionizing Strava charts2025-04-20T06:00:00+00:002025-04-20T06:00:00+00:00/2025/04/20/Productionizing-Strava-chartsI’ve been a Strava user for almost 15 years and have logged a lot of data there – something like 47,000 km across about 2,000 rides. Strava’s analytics are great for digging into a single ride, but I wanted to zoom out—see trends over time, track year-by-year totals, and watch how my average speed and power have changed.

So, during the pandemic, I built a Python app to do just that. It uses:

  • Bottle for handling web requests
  • Gunicorn as the dev web server
  • Requests for making calls to the Strava API (including OAuth)
  • pyChart.JS for rendering the charts

The first version was a simple, local-only tool that did exactly what I needed. It was a great experience building something from the ground up using the building blocks of well-designed libraries.

I’d always meant to make the app public but never got around to it—until recently. I decided to productionize the app as a learning exercise to get hands-on with some newer AWS tools. Here’s what I used:

  • OrbStack – to build and run Docker images for a clean dev environment
  • GitHub Actions – to automate Docker builds and deployment
  • AWS SAM – to orchestrate AWS services like ECR, Route53, API Gateway, and Lambda

This required a few code tweaks—handling Lambda event input and fixing some requests caching issues on a read-only filesystem—but most of the effort (and slow trial-and-error) was in setting up the CI/CD pipeline.

Along the way, I picked up a few things:

  • ChatGPT turned out to be a really helpful tool for ramping up quickly on new tech and syntax—it saved me a ton of time getting the initial AWS SAM and GitHub Actions configurations set up. That said, it tends to get confused when troubleshooting. Being able to clearly frame a specific question made a big difference.
  • AWS SAM seems really powerful, but it still feels a little magical to me, probably because I didn’t run into enough issues to force me to look under the hood.

The end result: a fairly streamlined build-and-deploy flow, and a live version of the app now running at https://stravacharts.3thirty.space*. The full source is available for anyone who wants to run it themselves (or using their own API key)

* Strava login support for users other than me is still pending approval 🤞

]]>Ethan SmithHello to github pages2024-11-03T21:19:12+00:002024-11-03T21:19:12+00:00/jekyll/update/2024/11/03/hello-to-github-pagesThis site (using jekyll via github pages) is replacing my 17-year old wordpress installation on 3thirty.net (archived on the wayback machine: https://web.archive.org/web/20240224095106/http://3thirty.net/

I’m not bothering to bring over any of the posts that are more than 10 years old, and are about very outdated technology, but hopefully I’ll have some more interesting things to add here soon.

In the meantime, old, outdated code is in my public github repos at https://www.github.com/3thirty

]]>Ethan SmithAlfred app for Flickr API documentation2018-08-17T06:00:00+00:002018-08-17T06:00:00+00:00/2018/08/17/Alfred-app-for-Flickr-API-documentationI have just published a simple Alfred App workflow that provides easier access to flickr’s API documentation and the API explorer. It’s not very complex, but is useful for someone (like me) who uses the flickr API regularly

You can get it from https://github.com/3thirty/alfred-flickr-api

Alfred App Screenshot

]]>Ethan Smith