iiinova — where your digital self keeps evolving

Reposted by Hadley Wickham

Dan Quintana @dsquintana.bsky.social · 25/07/2026

I just rebuilt my website using Quarto 🎉 My publication list generates itself from my Zotero library, and papers include links to materials and OA versions. The website's first post is on how to build your own, featuring an app that does all the coding for you: dsquintana.com/posts/quarto...

Screenshot of the Publications page, introducing the Behavioral Neuroendocrinology Lab and noting that preprints and materials are linked where available. Under a "2026" heading, five publication entries are listed with full citations (authors, year, title, journal, DOI links), each tagged with badges like "Data & code," "Preprint," or "Open access article."

Screenshot of the Talks page, noting that slides are posted on OSF. Under "2026," two talks are listed: "Recalibrating human oxytocin research: from methods to function to theory" (Harvard Interdisciplinary Oxytocin Research Initiative, online) and "Social media for academics" (NMBU, in-person), each with a "Slides" link. Under "2025," three more talks are listed: "Open science intro for psychology master students," "Start-up scholar workshop," and "Synthetic data workshop," all at the University of Oslo, in-person, each with a "Slides" link.

Screenshot of a blog post titled "How to create a simple academic website," tagged "Website" and "Quarto," authored by Daniel S. Quintana and published July 25, 2026. The text explains that the site was built using the Quarto publishing system, describes advantages over his previous R-based approach (easier Zotero-based publication updates, links to open data/code, and open-access versions of paywalled papers), and introduces two methods for readers to build a similar site: cloning the GitHub repo or using a Shiny web app that generates the code with the user's own info.

29514

Reposted by Hadley Wickham

Luis D. Verde Arregoitia @liomys.mx · 01/10/2026

🎨 formatted spreadsheets can now make the round trip in R. {forgts} reads a formatted .xlsx file into a gt table, and the new {gtxlsx} package by J.M. Garbuszus writes it back out with fills, fonts, and colors #rstats post available in English and Spanish luisdva.github.io/rstats/round...

luisdva.github.io

Round-tripping formatted spreadsheets with forgts and gtxlsx

From spreadsheet to gt and back again

0399

Reposted by Hadley Wickham

Hannah Shelley, MLIS (Maggots, Lycanthropes & Ice Skeletons) @hannahshelley.site · 02/10/2026

Please stop anthropomorphizing computers. They cannot "run" a program, because they don't have legs

392035283

Reposted by Hadley Wickham

Libby Heeren @libbyheeren.bsky.social · 02/10/2026

Nic ( @niccrane.bsky.social) taught us all about FREAKING PARQUET FILES on the #DSLab, which are wayyyyy cooler than I realized. Parquet files: Preserve data types Are fast & small Are column-oriented so you can read just 1 col if wanted! & Positron displays'em out-the-box! ▶️ youtu.be/2mQ4kc9J9uQ

youtu.be

What is a Parquet file anyway? Why is Parquet good? | Nic Crane | Data Science Lab

YouTube video by Posit PBC

56110

Reposted by Hadley Wickham

Carson Sievert (he/him) @cpsievert.bsky.social · 01/10/2026

New in querychat: multiple tables, chat history, data-dict, and easy extraction of findings into Quarto, Shiny, Jupyter, etc. querychat is now much more capable beyond just querying "pre-defined" Shiny outputs. Available for both #rstats and #python. opensource.posit.co/blog/2026-09...

The wizard for extracting findings into Quarto, Shiny, Jupiter, etc.

1329

Reposted by Hadley Wickham

Emi Tanaka @emitanaka.org · 30/09/2026

I’m pleased to announce that Vol 18 Issue 3 of The R Journal is out now! 🥳 Free & open access, R journal is made possible by our volunteer editors, reviewers + contributing authors. Thank you all! journal.r-project.org/issues/2026-3/ 🧵 20 articles 👇 #rstats journal.r-project.org/issues/2026-3/

journal.r-project.org

The R Journal: Volume 18/3

Articles published in the September 2026 issue

13811

Hadley Wickham @hadley.nz · 27/09/2026

This is going to save me so much time on slack

2140

Hadley Wickham @hadley.nz · 26/09/2026

Real pros skip the paper towels

4220

Reposted by Hadley Wickham

Victoria Rose Richards @aerialembroidery.bsky.social · 25/09/2026

'You'll choke too' - I wanted to reshare my wildfire aerial embroidery now summer is finally over as it really meant a lot to me. Autumn is here but as the fruit ripened much earlier than normal, I worry the little birds and animals won't have enough for the winter...

An embroidery hoop containing an embroidered aerial view landscape. The landscape depicts a burning wildfire, with dry forest on the right and smoke on the left, with a line of bright flames down the centre.

5434123

Reposted by Hadley Wickham

§ @sharons.site · 25/09/2026

blogpost-y version of the slides from my posit::conf 2026 talk last week: blog.sharons.site/what-context-do-your-ai-agents-actually-need i talk about not losing sight of building your own understanding and maintaining collaboration amidst the distracting and attention-absorbing AI landscape

091

Hadley Wickham @hadley.nz · 25/09/2026

He’s also super racist so there’s that

150

Hadley Wickham @hadley.nz · 25/09/2026

I don’t need to write that down, I’ll just remember it!

130

Hadley Wickham @hadley.nz · 24/09/2026

If you really want to procrastinate, now’s the time to switch to positron

241

Reposted by Hadley Wickham

Thomas Lin Pedersen @thomasp85.com · 24/09/2026

As announced at posit::conf(2026) #ggsql is now, with the release of version 0.5.0, in Beta. I'm so excited for the future of this project. Read all about what the latest version brings you in the release post

opensource.posit.co

ggsql 0.5.0: Readers, Writers, and Beta status

Announcing ggsql 0.5.0, with new readers and writers for getting data in and out, and the project's graduation to beta status.

14814

Reposted by Hadley Wickham

Etienne Bacher @etiennebacher.bsky.social · 22/09/2026

#rstats new blog post! "Snapshotting error messages", where I explain why I think all testthat::expect_error() should be replaced by testthat::expect_snapshot(error = TRUE), and where I present an easy, automated way to do this transition. www.etiennebacher.com/posts/2026-0...

etiennebacher.com

Snapshotting error messages | Etienne Bacher

Usually, when we develop software, we know very well how to use it and how it works. This is not true for external users, people who will make mistakes and won't use your software correctly on the fir...

1205

Hadley Wickham @hadley.nz · 22/09/2026

First CRAN release. Prior to that is gestation 🤣

161

Reposted by Hadley Wickham

Garrick Aden-Buie @grrrck.xyz · 21/09/2026

I had nearly the opposite experience at conf. AI has disrupted the way we work and changed so much about our jobs and our relationship with code. I've been in the same emotional space as Jon and the feeling is a real and valid and reasonable response to that change! And also...

1151

Hadley Wickham @hadley.nz · 21/09/2026

Possibly the best gift I ever received (from @winstonchang.bsky.social)

0172

Hadley Wickham @hadley.nz · 21/09/2026

I am a bad father: reshape turned 21 back on August 5th and I didn't even notice or celebrate #rstats

3522

Hadley Wickham @hadley.nz · 21/09/2026

It just feels like a genuinely new way of working to me, and it's a huge win for projects that need to get done where you don't care that much about the details. e.g. I found it a lot of fun for github.com/hadley/data-..., where the interesting stuff is more about the algorithms you pick.

150

Hadley Wickham @hadley.nz · 21/09/2026

I think it's worth noting you don't have to stick to one mode of working — some things benefit from careful craftsmanships and in other projects you want to bang out a quick prototype. (Also nothing about my current workflow feels like managing students.)

251

Hadley Wickham @hadley.nz · 21/09/2026

I still read 100% of the code I write for R packages, where the stakes are generally high. It's really important for testthat, dbplyr, roxygen2 etc, to continue to be robust high-quality code.

0120

Hadley Wickham @hadley.nz · 21/09/2026

I haven't read a line of reshape code in the last 10 years so I don't think AI is going to change that 🤣 But more seriously AI does make it substantially easier to handle the burden of CRAN maintenance, where you need to make a relatively unimportant change quickly.

130

Reposted by Hadley Wickham

Lucy D’Agostino McGowan @lucystats.bsky.social · 21/09/2026

🎙️ Today on Regression to the Mean Girls @sarahlotspeich.bsky.social & I chat with @hadley.nz about an over-engineered solution to purchase some beautiful art!

2163

Reposted by Hadley Wickham

Nic Crane @niccrane.bsky.social · 21/09/2026

ellmer 0.5.0 is now on CRAN! #rstats 🎉 New in this release: citations from web search, new ways to upload files, token counting before you send, streaming structured output, and AWS Bedrock Mantle support. Read all about it on the Posit Open Source blog: opensource.posit.co/blog/2026-09...

opensource.posit.co

ellmer 0.5.0

ellmer 0.5.0 is now on CRAN, with new ways to send files to the model, citations from web search tools, and new ways to update model price data.

0267

Hadley Wickham @hadley.nz · 21/09/2026

I have maintained the reshape package for 21 years. If that doesn’t convince you that I’m committed to long term maintenance I’m not sure what will.

3150

Hadley Wickham @hadley.nz · 20/09/2026

I think thinking like a software engineer is what caused this problem 🤣

3101

Hadley Wickham @hadley.nz · 20/09/2026

Would you be interested in capturing some of this in the ntile docs?

170

Hadley Wickham @hadley.nz · 18/09/2026

I think it was just copied from what SQL does

020

Reposted by Hadley Wickham

Data Visualization Society @datavizsociety.bsky.social · 17/09/2026

Data Craft Circle is back after a summer break! Come share your summer projects or what you hope to work on in the fall. Or if you've never joined us before, come hang out with others working on analog data projects! You can register for free at this link: buff.ly/AtfGRPU

Image promoting a September 21 event called Data Craft Circle, a space to bring your data encoding creativity whether you bead, collage, embroider, or just want to try something new.

0104

Reposted by Hadley Wickham

Jeremy Allen @jeremy-data.bsky.social · 17/09/2026

#RStats and #PyData peeps. Don’t fear IT at work. We’ll demo Package Manager (PPM) for your IT and their fear of the ecosystems will fade and open doors for you. PPM is your gateway to approved R and Python at work! Thanks to all at #positconf2026 who talked PPM with me. You got this!

031

Reposted by Hadley Wickham

Brian Helsel @bchelsel.bsky.social · 17/09/2026

I just finished my first paper written entirely in Quarto, and it has changed how I think about the manuscript-writing process. If you’re not an #RStats or #Quarto user, the biggest benefit is that you can automate repetitive workflows. Here are 5 things that made my writing process more efficient.

3465

Reposted by Hadley Wickham

Luis D. Verde Arregoitia @liomys.mx · 16/09/2026

Gave my talk "Can I offer you a nice IDE in this trying time?" at #positconf earlier today and it went well slides here: luisdva.github.io/rstats/posit...

4395

Hadley Wickham @hadley.nz · 16/09/2026

I assume the M is for Money

110

Reposted by Hadley Wickham

Kelly Bodwin @kellybodwin.com · 16/09/2026

This talk by @tigertang.bsky.social is my favorite of #positconf2026 so far, and is going in my "instant classics" pile: tigertang.org/posit_2026/ It's a super clear, actionable framework for how learning happens and how AI can be used to help not hinder. ✨

A slide with the text: imitate → Internalize → Transfer

A slide with the text: The Mastery Test

Can l
EXPLAIN IT? In my own words.

Can l
CHANGE IT?

When things change,
Can I use it
SOMEWHERE NEW?
On new problems.

2204

Hadley Wickham @hadley.nz · 16/09/2026

i.e. as data-dict.yaml evolves there will be a need for purely mechanical transformations, which in principle you could do with code, and I wouldn't want to inflict on a human.

000

Hadley Wickham @hadley.nz · 16/09/2026

Interesting. I see this specific application basically in the same vein as a code formatter; if data-dict was a bit further along it would do it automatically, but in the meantime you can use an LLM to simulate it.

100

Hadley Wickham @hadley.nz · 16/09/2026

I was suggesting using an LLM to keep up with a changing specification instead of making a human do annoying grunt work, which I think is a little different

110

Hadley Wickham @hadley.nz · 16/09/2026

How about this: if you use data-dict before it hits 1.0.0 I’ll personally fix any backwards incompatible changes that affect tidy Tuesday 😀

170

Hadley Wickham @hadley.nz · 16/09/2026

You just ask it to update the specific data-dict that’s out of date. Very minimal token usage.

100

Hadley Wickham @hadley.nz · 16/09/2026

5-10 years 🤣 but more seriously I think I’ll aim for something that has backward compatibility going forward by the end of the year. But also my experience so far is slightly out of data yaml file + detailed errors + LLM = quick fix

220

Reposted by Hadley Wickham

Tony Corke @matterofstats.bsky.social · 15/09/2026

Missing In Ways You Wouldn't Believe Missing, Presumed Dead Missing Without Leave Missing in Plain Sight Missing But Not Caring Missing Until the Auditors Show Up Missing Its Friends Missing But Drawing the Foul Missing Outside but Called a Strike

1143

Hadley Wickham @hadley.nz · 15/09/2026

Can we download all the data at once so we can analyse locally?

100

Reposted by Hadley Wickham

Jared Knowles @jaredknowles.com · 15/09/2026

Hey #positconf2026 here's a data source you might find useful to connect with or for your next training! If you want to learn more about the way public money has been used in your community over time - check it out!

1113

Reposted by Hadley Wickham

Simon P. Couch @simonpcouch.com · 15/09/2026

We're excited to share commons, an R and Python package that helps data scientists build trustworthy data agents for their organizations! Read more on the @posit.co open source blog: opensource.posit.co/blog/2026-09...

A screenshot of the blog post announcing commons. The hero image shows the hex sticker, a common kingfisher on a park bench.

14311