Yann Tourman 😎🚴🏻‍♀️

GitHub - iamYannC/r-podcast: A fun mini-project scraping the R-Weekly podcast A fun mini-project scraping the R-Weekly podcast. Contribute to iamYannC/r-podcast development by creating an account on GitHub.

github.com/iamyannc/r-p...

⭐

Who wants the RWeekly podcast on YouTube Music / Spotify?

#rweekly #rstats

Marvin panicks

yoo moltbook is wilddd

#moltbook #openclaw #moltbot #dontpanic

Is pak for the R community is essentially what uv is for Python?

#whitehat #scraping #nobodysreadingmyposts | Yann Cohen y’all think that’s harmless? Going through dozens of JSON objects (and sub-objects… manually! I enjoy this!!) sometimes reveals interesting data. This is just a simple scraper I ran through the night...

www.linkedin.com/posts/yann-c...

i think i found something that wasnt supposed to be found? idk. and why dosnt bsky allow private messaging?!

Did I say package? I meant dashboard 😉
Building the UX first, from the user perspective, then backfilling implementation feels like cheatin- n a good way. Jobs would've loved to be around in these great times.

#rstats #shiny

@thomasp85.com @juliasilge.com @gaborcsardi.org

As developers, we know our products inside out. Users don’t.
What feels obvious to us can be confusing enough to make people abandon our baby.
With agentic development, I can be that naive user. I interact with my product without knowing the internals.
I don’t even want to read my own docs

3/4 »»

I don’t write a single line of code manually.
I don’t even test anything until all features I imagined are implemented.

Once finished, I load_all() and use the package exactly as I envisioned it.

This is not TDD.
This is User Driven Development.

2/4 »»

My R package development has completely changed thanks to CLI tools and agents.
Automating docs and tests is old news. What I do now is different:

I act as my own user.

1/4 »»

Before I relaunched my bike trip across the Americas I toyed with scraping the transcripts and added plotting Ep' length to my cicd.

CICD broke long time ago but you can still see a clear trend :)
gonna be back at it soon!

github.com/iamYannC/r-p...
@rpodcast.bsky.social

Reading through well-written docs is much more pleasurable than blindly pasting LLM instructions.

Like we did in the old days, prior to Nov 2022

I'd take R any given day for any of these tasks. There's even an R implementation for PyTorch

Share your coolest alias with me.
Mine gotta be the 'shh' 😎

#linux #mint #linuxmint

Can someone explain this to me? Like, why? I'm in Martinique, it's part of FRANCE 🇪🇺🇫🇷🇲🇶

*Downloads are disabled, that's annoying. I'm paying YouTube

#youtubepremium
#vpn

I would never use this, or else I'd get to my destination too fast

OMG! Good ol' StackOverflow! Now thats something..! Both ChatGPT and Claude kept gaslighting me, saying my string was not unique enough and im like... sheeetttt

Then I GOOGLED and went through SO post and a guy fixed it by synch date and time - and it worked! and none of the llms suggested this

to make things extra clear, this is not gpt generated token! i tried randomizing myself and also used a token generator (which i also modified). im telling you, something is deeply broken here. it makes no sense, and if its trully is taken, they should come up with a better system

51m9g-6cc4l2-zplm-1999-z0k1bb-2n6fk9skg1jc7

if screenshot is cropped and you dont want to see the full
image, it says that this (^^) bucket name (along with MANY others i tried) is taken. yea right.

whats that bullshit?

AWS S3 bucket creation is BS. no way this is taken, and if so, the system is broken.

#aws #s3

I was fortunate to join as crew to a lovely captain, and together with another Belgium guy we sailed from French Guyana to Martinique, with a seven days pause in Tobago.

Sailing is no joke but I love the change. Now it's a short break from cycling, get some rest and be back on the road in no time

I think I'm gonna dig deep into plumber:: because I gotta get this amazing r logic into robust websites (love shiny but it ain't enough)

I'm in the Caribbean. Anyone here? Hitching & coding as times flies

Any future plans on adding S7 to this book?

It's my ground truth whenever I get confused but might be a bit outdated at this point?

@hadley.nz

I posted this as I was crafting and polishing my Data Scientist Gem in #Gemini Ai Pro (I think. The one year free student plan).

You can take a look and offer suggestions. LLMs (chatgpt and Gemini itself) approved 🔥💎

gemini.google.com/gem/1lyD7uYX...

Definitely letting my Gem know about this resource then. Thank you!

Also hunting all you's posit guys to add as instructions to read through the blogs, so thanks for replying 😎

That's totally new to me. Thank you. I might add it to my gem (that's how the idea came up to me in the first place..)

‎Google Gemini Meet Gemini, Google’s AI assistant. Get help with writing, planning, brainstorming, and more. Experience the power of generative AI.

Cool! I wasn't aware of it. Dang, my post got viral (3 comments!!) while I was crafting an R data scientist gem

I basically try to make it visit CRAN when I ask for things he's not sure how to execute.

P.s
#Gem written by me and polished using #Gemini itself.
gemini.google.com/gem/1lyD7uYX...

How great it would be if CRAN would get an AI version, basically feeding information on every package for AI to read, then decide if the specific package is worth exploring more (GH source, readme etc)
#rstats

No one cares about yet another epidemiology package and the global fish data set, for crying out loud.

That's it, no more shit posting about my beloved language. Take it or leave it

And even when there is a demo, it's always showing this boring EDA workflow. Oh look, it does its own scatter plot! Hooray. Seriously guys. It's dead. Why would anyone nowadays even bother starting learning R. And I've yet to mention the should-have-been-a-feature IDE that they can't get enough with

The R ecosystem is asleep, if not dead. YouTube is bombarded with Python tutorials for anything related to AI, from simple agentic workflows to cool mcp stuff and other services integrations. This really upsets me, as an R advocate (only one among a team of 7 Pythonists). 0 AI content with R.