Stats For Bios

I've heard that some certain annoying pollen has come early this year. Leicester has monitors or something.

Also, less combatorially, what are you hoping for that isn't on DAB?

You're entitled to your hopes and wishes, but if they ever force this apon me (with no off switch) I'm holding you accountable. I've spent the last 25 years doing my very best to avoid the talking in between songs.

How many hours have been spent writing rubrics that ultimately boil down to the totally subjective grading of:
A - excellent work
B - good work
C - ok work
D - just about ok
E - bad

Theory, without much basis in verifiable data.

Trump thinks Greenland is as big as it looks on Mercator projection. That's why he's excessively interested in it.

Book page with two dents.

I "invented" a "thing". For taking very rough, on the fly notes in a book. A folded corner and a couple of scores with my thumbnail points clearly enough to the bit I was interested in. I can often then remember what I thought about it.

Obviously has limitations.

"You might not have more courage, but you should be ashamed to show less." Iorek Byrnison

Interviewing for role of product tester for Ikea I take it?

The Irish National Liberation Army likes this.

BIG NEWS!

Minor, unimportant opinion. We should keep calling him Prince Andrew. If they had disowned him 10 years before stuff came out, maybe they can wash their hands. But they've been funding and protecting him. He's royal family and that familial shame still exists.

What could this look like? Some potential guidelines:

1. Try to use no more than 100 words in the whole poster (outside of text inside of tables and figures).

2. Make the figures giant.

3. Present *way less information.*

4. Ask a question on the poster to engage the audience.

Compute clusters. Peanut chocolate clusters. I might be veering off topic.

I've not even started trying to learn the future tense.
I'll just stick with Ir.
Lo que va a ser, va a ser.
Probably making every spanish speaker cry with how clunky it is.
Is it being? Or is it not being? There is a question.

The amount of time I spend on pages looking up conjugations of Ser and Estar is outrageous. Why won't it just stick in my head!

Yes definitely is sometimes (that was my postdoc work!) Some specifics are quite data rich, some are very data poor. And it's not at all obvious from the outside which is which.

For example, I reviewed a paper that fitted models to similar data and reported R^2 of 99% or something silly. Because, being estimates, all the noise was missing. Suddenly it's very easy to predict if it's already totally smoothed out!

Relating back to the original post, these are estimates. So fitting models to these for example doesn't make sense because they are already smoothed/biased etc. etc. And I don't think you can get the raw data.

I'm aware this data just might not be openly available. That's fine. But if so I want to work that out in half an hour, not in four days.

The fuzziness in usage between data and estimates is really annoying. I want to find some malaria mortality data to download and test some methods on. There is a thousand pages giving me malaria mortality estimates. I have to sift through all these to maybe find the ones with actual data.

Happy pancake day everyone. Just had a great lunch of lemon-only pancakes. The best kind.

Also the Agriculture album is incredible. But that wasn't the point of this post.

Been reading "you've not yet heard your favourite song" and thinking more carefully about streaming. The above summarises quite nicely the core of my thoughts. I only care about albums vs tracks vs playlists, or physical vs digital, or owned vs streamed insomuch as they affect that core.

"The focus needs to be on the listens that stand out not only as something to be recommended, but art that will build a space in your head which can be lived with and pondered well beyond the timeline’s attention deficit shelf life."
rcmndedlisten.com/2025/07/13/a...

We're pleased to have @betanalpha.bsky.social join us in Princeton tomorrow (Feb 17th) to talk about generative modeling!

Talk: 11am
Workshop (with Stan): 12:15pm

The zoom is open to public (see flyer).
@princeton.edu

Phwoar. Student spent two months collating data and coding up a complex, novel model. Seeing first estimates from reduced testing models today. Very exciting. This is what this job is about.

YouTube video by TLDR News EU What Went Wrong with Europe’s Pension System?

Interesting #DataViz choices here. Animation flipping between debt and budget multiple times.

I'm not totally convinced.

The url is timestamped.

www.youtube.com/watch?v=zFav...

This highlights important differences.
If the LaTeX code is wrong, it'll (nearly always) be obvious. Garbled table -> try again.

And ultimately, it doesn't matter in the same way. Nice docs are nice, but people don't die from misformated bullet point

If analysis code is wrong, it's not obvious.

Love ours. It doesn't fully replace doing it properly. But so much better than not at all!

If you need to run it three times in a room, do that. Run it two three times a week. It's a different thought process to manual.

"This is what it looks like when a simple clever idea is turning out to be too simple and insufficiently clever."
Glen McDonald

Pithy. I like it.