Ilya Kashnitsky

I've only been to one, 20 years celebration of R on 29 February 2020. I jumped in because it was in Copenhagen. And it was amazing 😍 (I bet years also an effect of COVID lockdown that happened two weeks later, and my next big on-site meeting was in July 2022)

In my experience, Gemini 2.5 Pro (free via AI Studio) gave best R results

Deeply honoured and genuinely surprised to receive this year’s European Demographer Award alongside @juliahellstrand.bsky.social and @fertdem.bsky.social 🥳Congratulations to all awardees 👏
Looking forward to seeing you in Berlin!

YES!!! @cosmostrozza.bsky.social for the ECR award 🏅For years and years Cosmo was doing a terrific job: producing great research, managing the big project after Jim's sudden passing, teaching EDSD students, teaching in the new CPop master program, supporting everyone around… Congratulations Cosmo! 🙌

Thank you!

Ah, I see so that each function help becomes a separate page, right?

What is the screenshot showing? It looks like some YAML listing but I don't know in which #rstats context one needs to list functions like this. Can you please explain?

Cool! I wonder if something similar exists for those who are not into pokemon lore?

Looks very interesting! It would be very useful to see these functions in action

🙈 ouch
But I bet in KZ it's horses

😍 beautiful #dataviz

I rarely come across examples of #dataviz that are awful and brilliant at the same time, use borderline unacceptable design techniques, and yet work near perfectly in delivering the key info 😅
🙌 well done @nurassyl.bsky.social, btw follow him for more cool stuff 🌱

Ronaldo and Messi are legends of such a scale that they became bigger than clubs they played for. Seaside souvenir shops are selling t-shirts featuring their career paths ⚽

July 2025 turned out to be my most running month so far, largely thanks to the gorgeous early morning runs on vacation in Croatia, for once I was really able to enjoy a 5k morning run daily for more than a week 💪👟🦿

🤣🤣🤣

Brilliant idea!
But isn't the classical problem of area size bias becoming critical here for representing any social variable?

@visnut.bsky.social @mjskay.com @jschoeley.com (tagging you because I remember your insightful talks at Rostock Retreat Visualizing Uncertainty)

I've committed this academic crime 🙈 have you? #AcademicSky

Yeah, this sounds too relevant for me. I would also not know where to start with a game. But that the power of LLMs — you can try something that looks only sketchily understandable

Google is testing a vibe-coding app called Opal | TechCrunch Google is testing a new vibe-coding tool called Opal, available in the U.S. through Google Labs, that lets users quickly spin up web apps with just a few prompts.

Google has just announced its vibe coding product. And seeing how much they struggle to win the userbase, they are likely to offer a generous free tier to start exploring, and their models are increasingly competitive
techcrunch.com/2025/07/25/g...

Google is testing a vibe-coding app called Opal | TechCrunch Google is testing a new vibe-coding tool called Opal, available in the U.S. through Google Labs, that lets users quickly spin up web apps with just a few prompts.

Google has just announced its vibe coding product. And seeing how much they struggle to win the userbase, they are likely to offer a generous free tier to start exploring, and their models are increasingly competitive
techcrunch.com/2025/07/25/g...

Sounds like a perfect (low risk!) pet-project to explore vibe coding 💡

Yes, the dataviz is iconic and beautiful. The story of the name is, unfortunately, dark as hell

Yes, I also enjoy using dark backgrounds, especially for the plots that use color and transparency heavily. I think we are too hardwired to think in terms of figures for academic papers, and those are bearing the conservative values of the publishing industry

To my big surprise I really liked one of your recent plots where you used a white smoother through a black jerky line, all with a grey background. I appreciated how the two lines were both present and separated visually.

Why do you think theme_grey didn't become a new norm in academic papers?

No! I didn't mean it to be anything reproachful, sorry if it felt this way. I was equally shocked to learn the story behind this popular band name

You know the story of ggridges, right?

Thank you for this thread! The idea to break everything is the best learning strategy for exploring the countless dataviz options. I picked up the idea years ago from @cedricscherer.com but it seems to be living on its own. You've pushed it to a new high though with the comparison to a proper plot😍

Oh yeah, the defaults are so powerful. (and ggplot2 defaults are a rare example of reasonable defaults 😍)

But do you think that theme_grey is the best one for exploratory analysis? Why isn't for example theme_minimal (I believe it's the usually the first pick in tweaking) chosen as the default?

A dynamic, interactive heatmap over a 3D globe?

It’s all possible with @mapbox.com and #rstats.

Learn how: walker-data.com/mapgl/articl...

Thank you for this 💛