Love and friendship (and data and government)

Good evening!

It’s been a great week. T and I had some great meals out, plus time with friends (also involving excellent food, including a delicious chocolate torte). I also had a good number of catch-up and “future thinking” chats with folks—thanks to all of you who make the time to help me think through whatever I ramble about with you, it really helps.

I also received a (only slightly overdue) ATI response this week, which, coupled with the fortuitously-timed launch of Observable Framework (on which more below), had me working on a dataviz project this week that I’m excited to share with you soon. It’s some deeply nerdy government stuff, but the kind that, if made more accessible to folks, can hopefully empower.

Okay, on to the links! T and I have ice cream, Oreos, and Schitt’s Creek to get to!

• “Observable 2.0” aka Observable Framework is a brilliant next step from the fine folks at Observable.
  • I’ve long been a fan of Observable, using it for data exploration and tinkering. Other than one timely calculator, though, I haven’t used it much for public projects—the easy editability of the interface, awesome though it is, has been a bit disorienting for folks when I’ve shared prototypes with them.
  • Framework, though, generates static sites powered by the Observable runtime, so you can make fast, well-formatted reports and dashboards that reliably display as you design them. It’s some clever stuff under the hood, and it’s been easy to get up and running. My two favourite parts are the ability to easily use non-JavaScript data sources (as I tend to use R for initial wrangling) and the static site output (so you can generate and deploy the report automatically on data refresh, using a CI / CD pipeline like GitHub Actions).
  • The Observable folks are also developing it in their classic personable style: the documentation says things like “we know this isn’t ideal, here’s the issue number if you’d like us to work on it”, effectively crowdsourcWen Schmidt for turning my attention to Observable—and JS for data work more generally](https://benschmidt.org/post/2020-01-15-webGPU/)—a while back.
• Found via Bostock’s announcement post, “File over app” by Steph Ango expresses an important philosophy for computing, to use the simplest possible file format, increasing the likelihood that it’ll endure.
• To continue the programming theme, Julia Evans crowdsourced and documented a host of useful git config options.
• Lorin Hochstein with two good ones this week: reviewing Trust in Numbers, describing the connection between the adoption of “quantitative rigour” in a field, and that field’s historical political strength (some strong fields have retained their “professional autonomy”, continuing to exercise judgement more often than they need to turn to numbers to make their case); reframing the questions to ask when things go wrong.
• Nisa on policy innovation and the “policy imaginary” is a welcome nudge (heh) to broaden thinking, nuance models, and embrace curiosity—a good reminder for public services that feel at times stuck in a deep winter.
• Amanda Clarke and Sean built on the Auditor General’s ArriveCAN report this week, reprising some of the findings and recommendations from their contracting research project. It’s been good to see the increased attention to and public discussion of contracting for government IT (as good as anything can be in the wake of something like the ArriveCAN story)—going forward, please, even more attention to these structural factors that cause folks working in government to turn to contracting arrangements like these.
• Finally, in a week that has been full of both love and friendship, Anne Helen Petersen in conversation with Rhaina Cohen on the same.

All the best for the week ahead!

Lucas