iiinova — where your digital self keeps evolving

Reposted by @datascienceweekly.bsky.social

Jo Wood @jwolondon.bsky.social · 24/09/2026

All that good work by Šavrič, Patterson and Jenny getting the UN to adopt the Equal Earth map projection, and here I am undoing it with Fourier harmonics. See also: observablehq.com/@jwolondon/t... and observablehq.com/@jwolondon/m...

A sketchy looking global map using the Equal Earth map projection. Land is shown in terracotta and sea in pale blue. Black coastal outlines are very rounded and are generated by extracting the 17th Fourier harmonic of each polygon ring and then applying Chaikin corner rounding. The goal is to allow any arbitrary TopoJSON to be transformed into loose sketchy geometry.

191

Reposted by @datascienceweekly.bsky.social

RLadies+ Bot @rladies-bot.bsky.social · 10/09/2026

📝 "Chart Checklist" Level up your data visuals with this expert guide. 👤 Helena Jambor (@helenajambor.bsky.social) 🔗 helenajamborwrites.netlify.app/post… #rladies #rstats #datavis #article #howto

011

Reposted by @datascienceweekly.bsky.social

Edouard Machery @edouardmachery.bsky.social · 03/09/2026

Well philosophy is willing to welcome all these undergrads

1122

Reposted by @datascienceweekly.bsky.social

MC Stan @mc-stan.org · 18/08/2026

There is now StanCon 2026 Uppsala playlist www.youtube.com/playlist?lis... At the moment talks by @mjskay.com, @aseyboldt.bsky.social, @paulbuerkner.com, @charlesm993.bsky.social are up, and we'll keep uploading more during the week

youtube.com

StanCon 2026 Uppsala - YouTube

23717

Reposted by @datascienceweekly.bsky.social

Matt Miller @thisismattmiller.com · 13/08/2026

Blog post about building bibliographic superwork clusters to help in catalog discovery. An experiment in using a local LLM to help label how books relate to each other: thisismattmiller.com/post/superwo...

thisismattmiller.com

Building Bibliographic Superwork Clusters for Discovery with Local LLMs

Judging work relationships in 88K clusters

064

Reposted by @datascienceweekly.bsky.social

Olivia Guest · Ολίβια Γκεστ @olivia.science · 09/07/2026

you 👏🏼 cannot 👏🏼 automatically 👏🏼 detect 👏🏼 plagiarism > the available detection tools are neither accurate nor reliable and have a main bias towards classifying the output as human-written rather than detecting AI-generated text link.springer.com/article/10.1... 1/

link.springer.com

Testing of detection tools for AI-generated text - International Journal for Educational Integrity

Recent advances in generative pre-trained transformer large language models have emphasised the potential risks of unfair use of artificial intelligence (AI) generated content in an academic environme...

518862

Reposted by @datascienceweekly.bsky.social

Al Merose @al.merose.com · 08/07/2026

I'd like to announce that at @openathena.ai, @jder.bsky.social and I helped @m2lines.bsky.social release Samudra 2. We scaled this neural ocean emulator to train on 16x the size of data in bytes on the same hardware budget. We can now skillfully predict 8 years of the ocean on a single GPU at a 1/4°

medium.com

🌊 Samudra 2: A Fast, Cheap AI Ocean Model, Now at the Scale That Matters

M²LInES’ neural ocean emulator now runs multi-year simulations at eddy-permitting resolution on a single GPU, turning a supercomputer-scale…

2626

Reposted by @datascienceweekly.bsky.social

Ryan Moulton @moultano.bsky.social · 25/06/2026

This post is another success of the heuristic "if you've written 10 tweets on a topic, flesh them out into a post on the real internet."

2833

Reposted by @datascienceweekly.bsky.social

Gnome Anne @annethegnome.bsky.social · 25/06/2026

Here it is! A digital archives resource list for the curious and nerdy 💚 I'll keep drafting my own essays and guides, but this is where I'll link people to external material. allmyfriendsarestories.neocities.org/AParchive/re...

allmyfriendsarestories.neocities.org

Resources

Resources related to digital archives.

13916

Reposted by @datascienceweekly.bsky.social

Alison Gopnik @alisongopnik.bsky.social · 17/06/2026

Second preprint, with Elizabeth Bonawitz, one of my favorite studies children search the hypothesis space more broadly and variably than adults. osf.io/preprints/ps...

osf.io

OSF

03310

Reposted by @datascienceweekly.bsky.social

Simon Späti 🏔️ @ssp.sh · 17/06/2026

I have been keeping a running timeline of data engineering acquisitions since 2022. It is getting hard to keep up. www.ssp.sh/brain/data-...

ssp.sh

Data Engineering Acquisitions (2022-2026)

Consolidation in the Data Engineering market is happening quickly. Tools from the Modern Data Stack get unified into bigger Data Platforms.

151

Reposted by @datascienceweekly.bsky.social

Simon Späti 🏔️ @ssp.sh · 11/06/2026

Many people dream of traveling to faraway places to experience new cultures and sights. Yes, you learn a ton, you learn new places, and the culture is hard to learn when not traveling. Travel locally, where you are.

ssp.sh

Travel Locally, Where You Are

Many people have the dream of traveling to faraway places to experience new culture and sightseeing.

371

Reposted by @datascienceweekly.bsky.social

Simon Späti 🏔️ @ssp.sh · 11/06/2026

Currently on Hacker News frontpage.

2131

Reposted by @datascienceweekly.bsky.social

Simon Späti 🏔️ @ssp.sh · 04/06/2026

Context is king, especially in the age of AI, where everyone is connecting their agents to their data stack. But writing SQL was never the hard part. Making it accurate and trustworthy against your warehouse has always been.

kaelio.com

Beyond the Semantic Layer: Building a Context Layer for the Agentic Era

A context layer puts your warehouse schema, joins, metric definitions, and business knowledge in one reviewable place so data agents query governed context instead of guessing field names. A look at h...

171

Reposted by @datascienceweekly.bsky.social

rmoff 🏃‍♂️🫖🥓 @rmoff.net · 04/06/2026

data folk: what's your goto open source data viz tool? is it still basically superset or metabase? (i just wanna point it at some data and throw some charts and maybe a map together; deffo not going the "just vibe something" route :-P )

341

Reposted by @datascienceweekly.bsky.social

Robbi Bishop-Taylor 🛰️🌏 @satellitesci.bsky.social · 14/05/2026

Happy to share some new #OpenAccess coastal remote sensing work published this week in @ijremotesensing.bsky.social! A new method for improving tide modelling accuracy by ranking and combining models using satellite data 🌊 🛰️ 📊 www.tandfonline.com/doi/full/10.... 🧵 Thread below!

21711

Reposted by @datascienceweekly.bsky.social

Simon Späti 🏔️ @ssp.sh · 14/05/2026

Amazing, thanks for having me. If you want to know more about technical blogging, I suggest subscribing to «Write that blog». My humble story and how I write is explained in this latest issue.

141

Reposted by @datascienceweekly.bsky.social

Cynthia Dunlop @cynthiadunlop.bsky.social · 14/05/2026

Our latest (and final, maybe) tech blogger interview takes you into the sticky web of @ssp.sh writethatblog.substack.com/p/simon-spat...

An excerpt from Simon on technical blogging

0112

Reposted by @datascienceweekly.bsky.social

Joanna Bryson @j2bryson.bsky.social · 13/05/2026

Same, except 3 decades.

18127

Reposted by @datascienceweekly.bsky.social

Quantian @quantian.bsky.social · 14/05/2026

US universities genuinely are like “welcome to Tennessee Agricultural College in Bumfuck Nowheresville. We invented the modern potato and our facilities include a particle accelerator, nuclear reactor, and TOP500 supercomputer. Tuition is $12,000 for instate students and we have 40,000 of them”

714233640

Reposted by @datascienceweekly.bsky.social

Giuseppe Sollazzo @puntofisso.bsky.social · 14/05/2026

Over the past 70 years of #Eurovision, love songs have long dominated. But in the past decade, lyrics about empowerment have become more common Will we see the big shift at #esc2026? Find out more at puntofisso.net/eurovision #ddj #ESC #esc2026 #dataviz

052

Reposted by @datascienceweekly.bsky.social

Nathan Lambert @natolambert.bsky.social · 07/05/2026

Visiting most of the leading Chinese AI labs, I'm struck by a culture that's extremely well suited to building LLMs with fewer resources, but one happening in a very different ecosystem, more companies at play, almost no data industry, etc. Full report: www.interconnects.ai/p/notes-from...

interconnects.ai

Notes from inside China's AI labs

Lessons from my trip to talk to most of the leading AI labs in China.

19514

Reposted by @datascienceweekly.bsky.social

Charlie Gao @shikokuchuo.net · 07/05/2026

OpenTelemetry has come to R 🎉 Shiny, plumber2, mirai, httr2, ellmer, knitr, testthat, and DBI are now instrumented: production-grade traces from R with no code changes. Set a few env vars and telemetry flows to the backend of your choice. Full write-up: opensource.posit.co/blog/2026-05... #Rstats

opensource.posit.co

Bringing OpenTelemetry to R in production

Posit has instrumented Shiny, plumber2, mirai, httr2, ellmer, knitr, testthat, and DBI with OpenTelemetry, and created tools for you to instrument your own packages, bringing production-grade observab...

27413

Reposted by @datascienceweekly.bsky.social

Berna Devezer @devezer.bsky.social · 30/04/2026

📣 The difference between replicable and not replicable is not itself scientifically replicable. 📣 New work with Erkan Buzbas, showing that verdicts such as "X% of results replicated" are based on an inferential machinery that doesn't work. arxiv.org/abs/2604.26268

Screenshot of a paper's title page. Title: "The Difference Between 'Replicable' and 'Not replicable' is not Itself Scientifically Replicable". Authors: Berna Devezer and Erkan O. Buzbas, both at the University of Idaho — Devezer in the Department of Business and the Institute for Modeling Collaboration and Innovation, Buzbas in the Department of Mathematics and Statistical Science. Authors contributed equally. Corresponding author: bdevezer@uidaho.edu.
Abstract: Replication studies estimate the replicability rate of scientific results by aggregating binary verdicts of experiments. Exact replications are rarely attainable, so most replication sequences are non-exact. Experiments differ in ways that matter and do not share a single common data-generating process. We formalize two statistical interpretations of this non-exactness. In a shared latent rate model (benchmark), experiments are exchangeable and depend on a common random replicability rate. In a conditionally independent rates model (operational), each experiment has its own replicability rate drawn independently from a population distribution. Under the shared latent rate model, even small variability among replicability rates induces an irreducible variance floor on the estimated mean replicability rate that cannot be eliminated by adding more replications. Under the conditionally independent rates model, the degree of non-exactness is not identifiable from standard replication data, because one binary verdict per experiment contains no information about between-experiment heterogeneity. Researchers therefore cannot tell which precision regime they are operating in or whether high- and low-replicability sequences can be distinguished in principle. As a result, the usual data structure of one binary verdict per experiment cannot support reliable demarcation between "replicable" and "not replicable" results and systematically understates uncertainty, making high- and low-replicability sequences appear discrim…

15545205