🌌 ¿Seguimos explorando el universo de R? @jeroenooms.bsky.social nos lleva en una segunda misión por R-universe de rOpenSci: APIs, herramientas y cómo hacer más visibles tus paquetes.
🗓️ Jue 12/11 · 10:00–12:00 · Auditorio 1
#LatinR2026 #rstats
Some stats about CRAN & Bioconductor:
There are 1918 packages with no package dependencies besides base R. There are 797 CRAN packages with dependencies on Bioconductor repos (3.11%), 424 to additional repos (1.66%). There are 2881 Bioconductor packages with dependencies on CRAN repos (77.89%).
Good, but let's centralize around a generic store, tabular byterefs with flexible UX and diverse workflows, not just one Rust lib and its python mono cult(ure)
https://www.earthmover.io/blog/ai-ready-scientific-data-archives
MICE always felt like magic to me 🪄 So I tried rebuilding it from scratch in R to understand it better. Lots of mistakes and help along the way, but my version ended up close to the mice package. Might be just luck. Still learning, but here are my messy notes! 🎃📊
https://www.kenkoonwong.com/blog/mice/
sf is great and fixed a lot of stuff and opened spatial up to many new people, it just didn't fix the things I cared about. That's all fixed now but you might as well yell on a mountain top about alternatives given the prominence of the Big Player.
I don't see how that's different for terrio/RustGIS, its servicing a new audience that was excluded one way or another. Same as Manifold.net, their market is a different pool, always was
I'm trying to couch the current crisis in many ways
I never liked how sf essentially captured all of the heart and mind of R spatial from 2017, the community has never recovered in a way I would like: as imo the foundations laid were way too narrow, sf really existed for itself and left no "shoulders"
I that kind of dismay now multiplied at the foundations level now - a nasty surprise and unprepared for, but inevitable that tech would go this way, but nothing inevitable about the framing
There are 62 new #rStats packages on CRAN:
- 17.74% are in English.
- 3.23% are in other languages than English.
- 0.00% use multiple languages.
- 79.03% do not declare any language.
ortho-cog-viewer gets a decent upgrade and UX design, now works better on mobile than it did
https://mdsumner.github.io/ortho-cog-viewer/
CRS-agnostic Cloud Optimized GeoTIFF and image or grid viewer with GPU-accelerated reprojection
The sign don't lie
5 💙 Todo este trabajo es posible gracias a mi Sovereign Tech Fellowship 2026 de
@sovtechfund , que apoya a quienes sostienen la infraestructura digital abierta y a sus comunidades.
Programa 👉 https://latinr.org/cronograma/programa/detallado.html
Si vas a estar en Medellín, ¡escribime y nos encontramos! #rOpenSci #RStats
4 💻 Viernes 13, 8:00–10:00: ¡HackeaR! Durante dos horas, en grupos chicos, vamos a hacer contribuciones reales a #RStats base y a proyectos de rOpenSci.
Si es tu primera contribución, vas a tener mentores y mentoras para que te acompañen
3 🎤 Jueves 12, bloque 10:00–12:00 (Auditorio 1): presentamos el currículo del Programa de Campeon(a|e)s de
@rOpenSci
(escrito con Maëlle Salmon, Paola Corrales y Alejandra Bellini)
Materiales abiertos en español, inglés y portugués.
2 🛠️ Miércoles 11, 14:00–17:30: tutorial “De idea a paquete: programá tu primer paquete de #RStats”.
Veremos: organización, código, documentación, tests y publicación. Discutiremos el rol de la #IA en el desarrollo de software.
¡Todavía hay cupos!
https://www.eventbrite.com.ar/e/1998690018649
1 ¡Nos vemos en #LatinR2026! #RStats
🎉 Del 11 al 13 de noviembre en Medellín voy a dar un tutorial, presentar una charla y co-organizar HackeaR, el hackathon de la conferencia.
serialize lazy xarray (loaded from open_mfdataset) to multi dimensional VRT:
https://pypi.org/project/xvrt/
(Open again faithfully as mdim with gdalxarray)
Here's another one
Oct 31 and so I was looking for the Misfits "Halloween" in my music library. I forgot I also have Lou Reed, Dead Kennedys as well.
I don't get software licensing at all, can't wrap Triangle for science because restricts commercial reuse (wtaf)
Embed MIT core in the most expensive incumbent desktop GIS with dopey Python workflows 🤔 ok no problem
Can't copy MIT to a port because a robot does it - that sounds like GPL-prescriptiv to me, oh and don't promote it "too much", do it quietly and be the single human user- no probs!.
Ship the MIT text for goodness sake, no one cared about liberal closed source reuse by corps 🥱
my cran watch, gives me some temporal perspective weirdly
https://mdsumner.github.io/my-cran-watch/releases.html
turns out "how many days since I submitted to CRAN?" is something I don't track well mentally and also care about #RStats
A lot of people dressed up like data centres here