UK Address Matcher logo
We are pleased to release `uk_address_matcher`, a free Python package for address matching and geocoding, developed by Tom Hepworth and me.
The package has several aims: simplicity, speed and accuracy.
Monochrome photograph featuring two women climbing on a rock face with ropes both wearing long dresses and hats
Rock climbers Lucy Smith and Pauline Rankin of the Ladies’ Scottish Climbing Club, 1908, Salisbury Crags, Scotland #WomensArt
Welcome to #WomensHistoryMonth
Shit, that sounds rough. I hope you have a swift recovery and a smooth journey back to exercise.
When D3 was released in February 2011, it transformed how we build data visualizations for the web and ushered in a new age of unbridled creativity in information design, data journalism, and beyond. Today, it remains the backbone of modern, interactive data viz.
Happy 15th birthday, D3 🎂
d3js.org
Choose your fighter. I am “Don’t trust NSE”!
bsky.app/profile/ding...
This is a terrific thread which really demonstrates the value of statistical reasoning.
bsky.app/profile/pwgt...
Just saw Leah W. share this on LinkedIn - great round-up of AI contribution policies! By @melissawm.bsky.social
github.com/melissawm/op...
1/8 So you're excited to go to the @ire.org 2026 NICAR data journalism conference!
And you want to follow all the speakers online. This tutorial will show you how do this quickly with #Rstats and @bsky.app.
schedules.ire.org/nicar-2026/#/
New: Drones are redrawing the map of war in Ukraine.
Relentless surveillance has pushed the battlefield 20km beyond the front, in both directions. Supplies arrive by drone; the wounded leave by robot.
This is the “kill zone” — and the future of warfare.
👉 ft.com/kill-zone
Yes! And as someone else just pointed out, the good natured humour of both threads is a reflection of what is good about the R community.
The thing I specifically remember from the evolution of JS was the move from promises to generators to async. I know someone who literally rewrote all of his packages to support generators using co and then did it again to support async.
But conversely I think the R community tends to move too slowly. I know plenty of people still using the magrittr pipe, map_dfr, httr etc with absolutely no curiosity to learn new ways of doing things. And I find that a bit frustrating.
The JavaScript community, for example, is like a hyperactive child that becomes immediately obsessed with every new toy, instantly forgetting how excited it was about the old toy, which it abandoned the minute the new toy came out.
I don’t want to weigh in on either of those things, but it got me thinking about how the cultural momentum to adopt new features varies across programming language communities.
Have seen two R threads this week debating language features: the first on base R dataframes vs tibbles; the second on the magrittr pipe vs the native pipe.
I think there is one way that technologists can future-proof their CV against this outcome, which is to start taking a serious interest in cybersecurity. Uncritical reliance on agentic AI is going to introduce so many vulnerabilities into vital business systems in the next five years.
It's one of those packages that once you find out about it, you wish you had found out much sooner!
My friend Ivan recently told me about this Python package, and if you need to routinely fetch data from a flaky API, it's a gift.
pypi.org/project/retry/
Title slide: Acting as if our research might matter - Analyzing data Image: A meteor about to strike the earth.
Clinical Researchers:
I am here to warn you about a meteor hurling towards us.
That meteor represents increasing scrutiny of clinical research; and in particular how we approach data collection, analysis, interpretation and reporting.
Prepare now, or face the consequences.
1/
love a bit of continvouclous morging
Also a classic!
Display an album that was important to you when you were nineteen.
bsky.app/profile/rsim...
A screenshot from the Mac application DaisyDisk showing the contents of the Application Support folder in my home directory. Slack is using 23.2GB.
Just discovered a 23GB Slack cache on my machine!
Short musings on "cognitive debt" - I'm seeing this in my own work, where excessive unreviewed AI-generated code leads me to lose a firm mental model of what I've built, which then makes it harder to confidently make future decisions simonwillison.net/2026/Feb/15/...
A special FT magazine issue all about maps is out today, w great pieces by @okr.bsky.social and @theboysmithy.ft.com, plus maps by me on how glacial melt is redrawing Alpine borders, the battle to redraw America and more
Read it online here (though best enjoyed in print!) www.ft.com/content/efab...
If you're a data journalist using Python and you are going to be at NICAR, we are running an introduction to Polars workshop.
schedules.ire.org/nicar-2026/#...
bsky.app/profile/emil...
If you are user of the Python package Pandas, I cannot recommend switching to Polars highly enough. #astrocode
I'm doing some groupbys for some plots, and a complicated set of string & numerical operations on *25 million rows* takes just 3 seconds on my laptop.
The Pandas equivalent: 2+ minutes. 🤯
I’m not saying this shouldn’t be an option but at least spell it out. All of the proposals to build a new Parliament in the North and turn the Palace of Westminster into a museum would be more expensive.
Okay. But the only way to save that money is to let a World Heritage site sink into the Thames. If you build a new Parliament elsewhere, you still have a failing building full of asbestos, ageing wiring, leaking pipes, falling masonry etc which you can’t use for any purpose in its current state.
For context, Polars is a modern, fast, elegant dataframe library for Python. We used it in all of our computational investigations last year, and in our live election results pipeline for the 2024 US elections.
