|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Rees-Mogg
3 replies 0 reposts 1 likes
|
Thomas Lumley@tslumley.bsky.social |
659 followers 320 following 1062 posts
Biostatistician. Baritone. Product of more than one country. May contain nuts. He/Him
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Rees-Mogg
3 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
So is the vampire lamppost going to lose his seat tonight?
1 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
They're really getting a recall because people are ending up in hospital and the company won't do a voluntary recall. The muscimol gives the FDA an excuse. I don't think anyone's convinced they understand the toxicity yet.
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
um. yay?
1 replies 0 reposts 0 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
A survey statistics shibboleth: multistage vs multiphase
notstatschat.rbind.io/2024/06/28/s...
0 replies 2 reposts 4 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
A survey statistics shibboleth: multistage vs multiphase
notstatschat.rbind.io/2024/06/28/s...
0 replies 2 reposts 4 likes
|
Andrew Gelman et al.
@statmodeling.bsky.social
[ View ] |
Interactive and Automated Data Analysis: thoughts from Di Cook, Hadley Wickham, Jessica Hullman, and others
statmodeling.stat.columbia.edu/2024/07/01/i...
0 replies 3 reposts 9 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I think xgboost or BART is still competitive for problems with lower signal to noise ratio and less training data.
0 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
What's brown and sticky? ...wait for it ... giggles A stick!
0 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I'd say "The Yeastie Boys" except that's already a brewery.
1 replies 0 reposts 0 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
If you want to use quizzes and bogus polls to "drive engagement" then go in good health. Just please don't try to turn the results into some sort of news.
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
...after a 50% budget cut..
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
"We've done enough work so far to be, I would say, 99.9 percent confident that we will get the right data going forward" --Papua New Guinea's Administrative Service Minister. I'd be nervous about saying that for *any* survey, let alone the PNG census
1 replies 0 reposts 3 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I think biostat is not as bad here: more people understand that the long-term point of theory is better data analysis. But it's still a problem.
0 replies 1 reposts 1 likes
|
grand theft eigenvalue 🔆
@akhilrao.bsky.social
[ View ] |
📉📈 the engineering piece is actually huge here Not infrequently the really frontier methods and so forth need to be implemented by people who can (a) use lots of compute, (b) think carefully about the math, or (c) do some combination We don't support such folks in econ. Our loss
1 replies 2 reposts 5 likes
|
ᴅᴀɴɪᴇʟ ᴍɪʟʟɪᴍᴇᴛ
@dlmillimet.bsky.social
[ View ] |
#EconSky 📉📈
I'm back at it and have a new blog post on #EconometricSausage
It is called "The Great Divide"
dlm-econometrics.blogspot.com/2024/06/the-...
Thanks for reading and circulating! I hope people will think seriously about it. 🙏
3 replies 11 reposts 25 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
No person shall...make a false or misleading representation that goods are new, or that they are reconditioned, or that they were manufactured, produced, processed, or reconditioned at a particular time
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
1 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
There would also be fish called Stabby and Bitey except there was too much competition
1 replies 1 reposts 4 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Call me Ishmael -> Call me, Ishmael
2 replies 0 reposts 4 likes
|
Derek Lowe
@dereklowe.bsky.social
[ View ] |
Let’s do some flowers! All of these are from the home gardens here. First, a ground clematis
2 replies 2 reposts 16 likes
|
Eddie Clark
@dreddieclark.bsky.social
[ View ] |
Let's do something a bit different for my usual nerdy SFF lists. For pride month I'll do 1 queer SFF book a day from each of the past 30 years. Will add 1 to the thread each day (and 2 today I guess because I'm 1 day behind).
4 replies 45 reposts 126 likes
|
Mike Dickison
@adzebill.bsky.social
[ View ] |
It’s a shame the NZETC archive will be retaining the incorrectly-applied Creative Commons licenses: for example this letter by Katherine Mansfield’s copyright expired many years ago, but it still has Attribution Share-Alike restrictions. nzetc.victoria.ac.nz/tm/scholarly...
1 replies 2 reposts 6 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
And in mice without adaptive immune systems. For good reason, but still.
0 replies 0 reposts 0 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
The open-access paper is linked from the press release
0 replies 0 reposts 0 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
It's a good book. More people should read it
0 replies 0 reposts 1 likes
|
Mark Baxter
@aotearoabadger.bsky.social
[ View ] |
A reminder that a phrase ending "with bells on" is an ancient curse involving being infested with Morris Dancers.
3 replies 1 reposts 7 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Peter Ellis' bayesian state-space models
freerangestats.info/elections/nz...
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
At the first scientific conference I attended, one of the international speakers had been to the Great Barrier Reef the previous week, and it showed. He was presenting on melanoma treatment 😬
0 replies 1 reposts 3 likes
|
Rob Hyndman
@robjhyndman.bsky.social
[ View ] |
An update to the fpp3 package for #rstats robjhyndman.com/hyndsight/fp...
Lots of new data sets
0 replies 4 reposts 4 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Pretty comprehensive ban, eh
#PunctuationMatters
0 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
There are literally paintings from ancient Rome in Pompeii! With people in them! In colour!
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I wrote a thing about the difference between using the true value and the estimate in a sandwich estimator #statsky
notstatschat.rbind.io/2024/06/27/e...
1 replies 1 reposts 3 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
The main time that you really *need* the full multistage variance formula is if you're writing software that you want survey people to take seriously.
0 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I handed out a bunch of Certificates of Permission to Write Badly to graduate students (with the stipulation that you then need to revise and get edits from your supervisor) [Idea taken from James Macdonald]
0 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
The 'cheese' option is specifically referencing Patrick Heagerty
0 replies 0 reposts 0 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
I wrote a thing about the difference between using the true value and the estimate in a sandwich estimator #statsky
notstatschat.rbind.io/2024/06/27/e...
1 replies 1 reposts 3 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
This is nicely ambiguous between being worried that people think deranged sex-cult conspiracy is natural and being worried that they think avoiding it is is natural (rather than needing care and effort)
0 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
"There were crimson roses on the bench; they looked like splashes of blood" -- Strong Poison, D L Sayers (The last chapter begins "There were golden chrysanthemums on the judge's bench; they looked like burning banners")
0 replies 1 reposts 5 likes
|
Kevin Collins
@kwcollins.bsky.social
[ View ] |
This is not a good question format. And I know the people at the Times know the research on this
1 replies 4 reposts 14 likes
|
Gerry Canavan
@gerrycanavan.bsky.social
[ View ] |
attn scholarly organizations, this is now considered best practice
16 replies 354 reposts 1180 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Ask them to name three dozen types of potato
0 replies 0 reposts 1 likes
|
Julia M. Rohrer
@dingdingpeng.the100.ci
[ View ] |
If an effect falls in a forest and no one is there to determine the mechanism, is it even causal? New post in which I try to clarify some things--claims about causal effects are indifferent to mechanisms; heterogeneity does not invalidate average estimates. www.the100.ci/2024/06/26/s...
8 replies 24 reposts 50 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Is this you in this recording? And were you lying then or are you lying now?
1 replies 0 reposts 2 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Do you know the Flanders & Swan song?
youtu.be/CrcLhG1madM?...
1 replies 0 reposts 1 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
Back in the Stone Age, when capital letters sorted ahead of lowercase, it made more sense. Hence also 00README
0 replies 0 reposts 4 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
The FAQ for the NZ Garden Bird survey has some great explanations of design principles
gardenbirdsurvey.nz/faq/
1 replies 2 reposts 7 likes
|
fooeynet
@fooeynet.bsky.social
[ View ] |
6 replies 5 reposts 38 likes
|
SpecsChic
@specschic.bsky.social
[ View ] |
Hey AoNZ, want an excuse to spend an hour where you can contribute to science just by sitting and observing your own garden? It's annual garden birds survey time 29 June – 7 July 2024. All you need to know can be found here.
gardenbirdsurvey.nz
0 replies 37 reposts 35 likes
|
Thomas Lumley
@tslumley.bsky.social
[ View ] |
"the oriflamme, which might only be unfurled in a holy cause, was taken up against the Flemish because they were urbanists, that is, infidels"
#SlightlyOutOfContext
0 replies 0 reposts 1 likes