Git Push Origin Me
My personal website has been quietly embarrassing me since 2021. Not in a dramatic way — it was still up, still rendering — but frozen in time in a way that felt increasingly dishonest. The version of me it described was a dissertator with a vague set of interests and a handful of blog posts from a pandemic year. I’ve since graduated, changed roles, became a hobby potter, and taken on a fair number of other things the old site didn’t know about.
The deeper problem is that I am, professionally and creatively, several people at once. I work as a research Data Science consultant for the University of Wisconsin-Madison, main a website to showcase my pottery, maintain a second website for the aikido studio, and a GitHub Pages site for the R community I’m starting in the fall. None of these talk to each other. If I write something at the intersection of pottery and data science — which happens more than you’d think — I have no clean way to surface it in both places. Every post either lives on one platform or gets manually duplicated, which means it usually doesn’t get written at all.
The need was clear: a unified personal site that serves as the canonical source of truth for everything I write, with a way to route content outward to the right platforms automatically.
The hurdles were predictable. The old site ran on Hugo, which I had stopped understanding some time ago. The Netlify project existed but was disconnected from its GitHub repo — apparently I unlinked them at some point and forgot. The platforms I write for speak different languages: some take Git pushes, some have REST APIs, one has a Content API that may or may not be accessible on my plan. Getting all of this to play nicely with a single source file was going to take some design work before any coding.
The solution I landed on is a hub-and-spoke model. This site is the hub, built in Quarto and deployed via Netlify from a GitHub repository. Every post lives here as a single .qmd file with a YAML header that carries routing metadata: which platforms it should go to, and whether it’s ready to publish. A small R script reads that metadata and handles dispatch — a git push for static sites, authenticated API calls for the others.
The dispatch layer is still being built. But the hub is up, the pipeline is running, and the first posts are in. It’s a start.