#rstats Any sparklyr / Kubernetes experts able to help me with this issue? github.com/sparklyr/spa... π
05.12.2025 16:10 β π 1 π 3 π¬ 0 π 0@gmcd.bsky.social
Economics and data science | Former academic now working in the tech sector | Views my own | πΏπ¦ in πΊπΈ | https://grantmcdermott.com/
#rstats Any sparklyr / Kubernetes experts able to help me with this issue? github.com/sparklyr/spa... π
05.12.2025 16:10 β π 1 π 3 π¬ 0 π 0To rephrase: there's a load of unrealised potential for GPU metrics, which is what Apoorva's post shows (esp. on the structural and Bayesian side). But if you want a faster lm / linear model, you're not going to beat .lm.fit.
05.12.2025 04:33 β π 5 π 0 π¬ 1 π 0Caveat emptor.
I know you aren't claiming otherwise, but ofc the big caveat is that you're benchmarking against torch CPU... not a standard, optimized max.lik. library. It's _very_ hard to get GPU improvements on closed form algos & that's before you a/c for the I/O to your graphics card.
Please enjoy this blog post based on my colleague Kyle Jensen's talk surveying different approaches to AI coding abilities
paulgp.com/ai-coding/20...
A houseplant casts a shadow of heart-shaped climbers in the background
Hearts on the vine?
29.11.2025 03:48 β π 4 π 0 π¬ 0 π 0Does setting one of the "dynamic" themes give you what you want? grantmcdermott.com/tinyplot/vig...
28.11.2025 18:57 β π 1 π 0 π¬ 0 π 0Background/presentation, ICYMI. bsky.app/profile/gmcd...
27.11.2025 17:31 β π 1 π 0 π¬ 0 π 0Example of dodged pointrange (tiny)plot, with aligned ribbons layered on top.
Happy `tinyplot` v0.6.0 (codename "Thanksgiving") release day to all those that celebrate. Some new features, but mostly bug fixes and internal improvements.
grantmcdermott.com/tinyplot/NEW...
#rstats
This will be a decidedly unpopular Springbok view, but SA's level of scrum dominance can't be good for the game. I don't know how to "fix" it, but I don't think it's just a question of a better bench. It's starting from the off.
22.11.2025 20:16 β π 2 π 0 π¬ 1 π 0Faz must have been apoplectic in the halftime change room.
First 20 min and setpiece aside, we were pretty poor TBH. Obviously, that's a reflection of the sheer grit from your lads. But I'm struggling to remember the last time SA played with such a lack of ambition.
Extraordinary. Doris MoM by miles for me.
22.11.2025 19:47 β π 1 π 0 π¬ 1 π 0It's funny how this level of set piece dominance can ruin the game for both sides. I can't remember when SA last played with this lack of ambition.
22.11.2025 19:34 β π 1 π 0 π¬ 0 π 0What is the Irish word incredible courage combined with extreme stupidity.
22.11.2025 18:40 β π 0 π 0 π¬ 0 π 0First 20 min of this game almost as intense as @kjhealy.co's posit::conf keynote.
22.11.2025 18:09 β π 0 π 0 π¬ 0 π 0"Great news, I booked a weekend away."
<Zen master: we'll see>
"Bad news. We have to miss Springboks v France."
<ZM: we'll see>
"Good news. The Boks get a first half red card, so we miss that at least."
<ZM: we'll see>
"So, about the actual result..."
Zen master vibes as my farsighted wife booked family travel weeks ago, causing me to miss the second half. #FRAvRSA
08.11.2025 21:17 β π 2 π 0 π¬ 0 π 1A bloodied Damian McKenzie lining up a matching winning penalty against Scotland
Smile before the kick
The non rugby mind cannot comprehend these two images.
08.11.2025 17:05 β π 5 π 1 π¬ 1 π 0π¨ How to create elegant #dataviz with base #rstats and the #tinyplot pkg
π useR! 2025 presentation by Grant McDermott @gmcd.bsky.social
www.youtube.com/watch?v=MOB-...
Huge JVDF fan tbh.
Minor 2024/25 form blip aside, and he looked back to his best over the weekend, just world class every game.
<insert bad joke about him finally going through the gate>
02.11.2025 21:39 β π 1 π 0 π¬ 1 π 0Beyond dispute IMO.
And it's funny to see people twist themselves into knots trying to argue otherwise--including far too many of my fellow πΏπ¦ countrymen--for precisely the reason that Squidge points out: we all instinctively *feel* that they are the greatest.
Thanks! I'm unreasonably pleased with it tbqh. Would dig to get your feedback.
Lots of additional material on the website. grantmcdermott.com/tinyplot/
I gave this talk on an empty stomach and about three hours sleep. So I recommend watching it >1x playing speed...
(Also: please have pity on the tall people in your life, forced to stare directly into projector lights.)
Nice. Looks like my `tinyplot` talk from userR! 2025 is up on YouTube. youtu.be/MOB-QS-vMeY?...
#rstats
Hmmm, I wouldn't say the problem is the availability of binary-build channels. Alongside r2u (great albeit Ubuntu only), there's rspm and the Posit manylinux repo (both multi-distro). The problem is conda hijacking and/or overriding these channels, so that they don't work like normal :-/
03.10.2025 03:27 β π 0 π 0 π¬ 1 π 0Yeah, unfortunately the inconsistency is the killer here. I've also run into a (related?) problem of missing pre-compiled binaries on some Linux distros that we use at work. So you end up having to install + compile everything from source β οΈ
02.10.2025 21:35 β π 0 π 0 π¬ 1 π 0RE: lack of love for conda-forge, see this thread by @jeroenooms.bsky.social fosstodon.org/@jeroenooms/...
02.10.2025 21:00 β π 1 π 0 π¬ 1 π 0It's a constant stream of untaxed negative externalities that I can only assume are part of some cunning grand plan for language hegemony.
30.09.2025 18:53 β π 1 π 0 π¬ 1 π 0Old ground, but I still cannot believe how poorly π plays with other languages. Okay, we're going to force you to do everything in a venv (b/c otherwise we break everything...) But that venv is going to auto instantiate and override any system level env vars or processes that you've set β οΈ
30.09.2025 18:51 β π 3 π 0 π¬ 1 π 0Fair points. Still, if you feel like modifying / playing around... bsky.app/profile/gmcd...
30.09.2025 00:14 β π 1 π 0 π¬ 0 π 0