Just released in v0.2.0 on alpha.cartokit.dev—a reimagined interface for viewing (and sorting) your data's attributes alongside cartokit's generated program! This makes it much easier to view multiple columns simultaneously, while giving maximum vertical space to the code.

A tiny bit of Friday UI transition🧈 coming to the next patch release of alpha.cartokit.dev. Now when you open the program view and data view, floating elements on the map interface adjust to stay visible. Keep your program, data, and tools right where you need 'em!

The problem with writing code with your fancy noise-canceling headphones on is that you can't hear your MacBook fans whirring from the MASSIVE memory leak you just wrote 😅

Exciting news in cartokit—we now use DuckDB (via duckdb-wasm) as our backing data store! This enables exciting things like reprojection and buffering (via GEOS!) that run 𝘧𝘶𝘭𝘭𝘺 𝘰𝘯 𝘵𝘩𝘦 𝘤𝘭𝘪𝘦𝘯𝘵. 11k points being reprojected and buffered 𝘪𝘯 𝘵𝘩𝘦 𝘣𝘳𝘰𝘸𝘴𝘦𝘳!

Rinsing and repeating today, but for OOPSLA this time 😅 This album is so damn gooooooooood.

cartokit v0.2.2 is out with a big set of UI improvements around color scales, transformations, and basemaps! More excitingly, we've started adding Playwright tests to replicate, automate, and verify our cartographic design workflows. Get Mapbox GL JS programs in <20s!

cartokit v0.3.0 is out! Alongside sweating a lot of UI details, we now support exporting transformed data alongside Mapbox GL JS code. Here we grab 2000 wildfire perimeters from the NIFC's ArcGIS API, grab their centroids for a proportional symbol map, and export the GeoJSON!

📢Jupyter + Python folks 📢 My pal Eric Rawn is running a (paid 🫰) user study researching the present and future of computational notebooks. Help him help you have a better experience with notebooks! Info: github.com/erawn/pagebrea… Signup: docs.google.com/forms/d/e/1FAI…

It was such a privilege (and so much dang fun!) getting to work on maps, graphics, data analysis, and cloud infra with the folks grist on this piece. Chuffed that we won this award and even happier that this amazing piece of reporting is getting the attention it deserves 🙌

cartokit v0.5.0 is out with a big new feature—channels! Channels allow you to map multiple attributes to different visual channels. Here we take 10k powerplants, size them by MW output, and color them by energy source. This was a 2017 WaPo graphic, made in <40 secs in cartokit!

It was deeply meaningful (and fun!) to develop the interactives and data analysis for this piece alongside Clayton Page Aldern, Tristan Ahtone, Maria Parazo Rose, and many others grist and High Country News. I'm so proud of this work and how it's revealing the signal in these datasets.

A few quick thoughts on the A's and the utter wickedness at play. Pro sports rely on the emotional connection between team and fan. It's a pact. The fan buys in, caring deeply about 26 souls wearing the same laundry. The fan makes it all matter, makes it mean something.

You're probably gonna see a lot of maps tonight. If you happen to need to program one, we've got you covered at alpha.cartokit.dev! Get started here: docs.cartokit.dev/get-started/ (Oh, and here's the TX 2020 Senate vote breakdown by county, in 39 seconds.)

I am _begging_ newsrooms to stop using non-population-normalized categorical choropleths to show the incoming vote share 😭

Flitting off to 🦋 Find me over there for more on maps, graphics, web dev, PL, (occasionally) baseball, et al. bsky.app/profile/parkie…