Mattan S. Ben-Shachar 🎗️🇮🇱🇺🇦

Output of dex_sentence function. "Elegant shellder uses wrap greedily" "Well-documented golbat uses lovely kiss sedately" "Old-fashioned gyarados uses disable cautiously" "Impressionable gastly uses lovely kiss scarily" "Tidy graveler uses amnesia questionably"

Output of dexid function "heavenly_tangela" "bite-sized_krabby" "frizzy_machamp" "impeccable_diglett" "firm_spearow" "delightful_ditto" "complete_golem" "weary_pidgeot" "winged_bellsprout" "realistic_machop"

Hex sticker of the R package dexid with showing a non existing pokemon that looks like a cross breed of charmander and eevee

If you are in need of memorable/human readable uuids, may I suggest {{dexid}}, an #rstats package that creates ids from adjectives and pokemons? 😀

I've created a .here file and it works great 😃

Academia is basically a collection of people who got lucky early on and mistook it for genius. doi.org/10.1073/pnas...

Add support for afex::lmer_alt · Issue #844 · easystats/performance See https://sites.google.com/view/jamiecockcroft/blog#h.bv1nxxxwvxwt

I can't seem to replicate lmer_alt not working. If you have a reprex, please post it

github.com/easystats/pe...

I ran an #rstats lab for many years before our foray into #pydata, so wouldn’t say my experience is super generalizable. But going back to play with R and seeing improvements to wasm and shinylive, ellmer, ragnar, vitals, and the entire tidy ecosystem reminded me just how batteries-included it is.

I’m much more comfortable with vibe coding (if needed) for a functional language with piping, where everyone agrees on which objects are required, and there’s a lower chance of bad design decisions being made. The resulting code is much less likely to need dramatic refactoring.

I should note that all unis here require STEM majors to take some nom-STEM courses to graduate: social sci, humanities, art, sports... I suspect 99% of engineers took tennis or something...

I think that in the coming year I'll be using Positron in class, but in my intro to R class I will still be teaching w/ RStudio (but hope to also switch there in the following year).

Sidenote: My BA is a double major in psych, socio-anthropology, and I learned so much about history, economics, society, war, ....
I'm not saying everyone needs to do *that*, but STEM folk seem to me to lack an education in a certain kind of critical thinking that is more common in the humanities.

And the consequences are clear: social, moral, philosophical issues are often discussed as engineering problems (cost, profit, effort...)
because people lack the mental tools or historical context to think of these issues in other, more productive (but not in the capitalist sense) ways.

I don't think it morally corrupts to study STEM. But..

I live in a country that invests in having STEM in school at the expense of humanities (sometimes literally); over the yrs matriculation requirements in lit, hist, civics have been reduced with pressure to take advance STEM classes instead.

"I asked the AI" is not research You can't reliably fact-check or research obscure topics by asking an AI. It may know less about the topic than you do.

"I asked the AI" is not research.
clauswilke.substack.com/p/i-asked-th...

This workshop is going to be great fun, and I really look forward to meeting everyone!

It would mean a lot to me if you signed up!

Oh face spoilers! I assumed you *were* a hippie minifig!

captain america is holding a hammer and says avengers assemble Alt: Captain America holding Mjölnir and says: “Avengers… Assemble!”

Mixed-effects modellers, assemble!

Just dropped a (very niche) blog on the weird quirks that occur with uncorrelated random effects and the trifecta of packages: lme4, afex, and performance (in R).

Read here 👉 sites.google.com/view/jamieco...

#rstats #MixedModels #lme4 #afex #performanceR

Hmmm we ( @easystats.github.io ) definitely should support lmer_alt.

Can you add some info on issues you've encountered?
github.com/easystats/pe...

If you have infinite data and compute, 1-nearest neighbors is also optimal.

It's a very frequentist way of thinking. Bayesians de-emphasize estimators.

Would 100% watch a Trad Wife talking about Fisher vs Pearson-Neyman while churning butter.

Maybe I'm also misunderstanding 🤷‍♂️

I agree that this statement is true about Trad Stats™️.

But I would say the there are few people who do Trad Stats™️ professionally.

Between ML™️ and Trad Stats™️ there is a whole chasm of tools that can be applied and learned and taught and expanded on.

Last time I used py in vs was like 10 years ago, so maybe?

It works quite well 😀
But it's very inefficient, time wise.

I would like true posterior distributions, but could never quite figure it out.

Using a .here file with my position projects is working great 😃

I found this convinced me even more to stick with ROC. Having a metric sensitive to the base rate seems counterproductive?

Hmm why "but"?

I really should revisit that function. I feel like I would have implemented it differently now.

#Position: one small step for #rstats, but one giant leap for #python.

(Really, how are people doing data analysis in python with other IDEs?)

Introducing the Data Explorer in Positron!

Quickly view raw data files (CSV, Parquet, etc.) or dataframes from your #Python / #RStats sessions with a data grid, summary panel, and filter bar.

Learn more: positron.posit.co/data-explore... #Positron

Solved it -
ctrl+shit+p > spellright: select dictionary

(no languages were selected)