Peter Ellis

@peter_ellis
605 Followers
519 Following
626 Posts
Director, Statistics for Development Division, at the Pacific Community. Tweets only represent my views, and are most likely to be about data and #rstats, maybe with bits of history, meta science and philosophy thrown in.
Tag1#rstats
Tag2#pacific
Work websitehttps://sdd.spc.int/
Personal bloghttps://freerangestats.info
Peter EllisDec 23, 2024

I blogged about the interesting link between #mentalhealth - depression - and voting for Trump in the 2024 election. A relationship between these variables at the county level stands up, in a model correcting for spatial autocorrelation and a state level random effect (using #rstats mgcv::gam).

Of course the causality is unlikely to be depression -> vote GOP, but to reflect an underlying common economic or social cause.

https://freerangestats.info/blog/2024/12/23/depression-and-vote

Depression incidence by county and vote for Trump

Multi-level modelling with spatial auto-correlation! I look at county level data on incidence of depression in 2020, and voting for Trump in the 2024 US Presidential election, and conclude that there's something there, but of course there are lots of potential explanations of what is behind the relationship.

free range statistics
Peter EllisNov 15, 2024
Tyler Morgan-WallAug 16, 2024

Fantastic work and journalism by the team at Bloomberg Graphics that used R, rayshader, and rayrender to produce the 3D animations to accompany this story on the rubble and scale of the cleanup effort in Gaza!

"Gaza Reduced to 42 Million Tonnes of Rubble. What Will It Take to Rebuild?"

By Fadwa Hodali, Fares Akram, Jason Kao, Jennah Haque, Jeremy C. F. Lin

Link: https://www.bloomberg.com/graphics/2024-gaza-who-will-pay-to-rebuild/

#RStats #rayshader #rayrender #dataviz

Who Will Pay to Rebuild Gaza After the Israel-Hamas War?

As US-backed cease-fire talks restart in Doha, calls to reconstruct the Strip are becoming louder

Bloomberg
Peter EllisNov 15, 2024
At #UNDataForum in Colombia and, as is common at large events, the plenaries have sign language interpretation. Are sign languages comprehended internationally, or would only Spanish Sign Language users understand this?
Peter EllisNov 14, 2024
Thomas LumleyNov 14, 2024
Better displays vs better visualisations in stats #rstats
https://notstatschat.rbind.io/2024/11/14/brute-force-and-ignorance/
Brute force and ignorance - Biased and Inefficient

Peter EllisNov 13, 2024
Nick TierneyNov 13, 2024

Super stoked to announce that {greta} version 0.5.0 is now on CRAN! Thank you to Nick Golding for making this release happen, and for his continued support, I learnt a lot!

Some features I'm really excited about:

- Now uses Tensorflow 2.0 (which involved a heavy refactor of internals)
- Installation of python dependencies has been revamped - it is closer than ever to "just working".
- Some new print methods for greta arrays and MCMC, these will only get better.

You can read the full news below:

https://github.com/greta-dev/greta/releases/tag/v0.5.0

#rstats

Release greta 0.5.0 · greta-dev/greta

This version of greta uses Tensorflow 2.0.0, which comes with it a host of new very exciting features! Optimizers The latest interface to optimizers in tensorflow are now used, these changes are de...

GitHub
Peter EllisNov 12, 2024
Can a Stata or survey expert suggest why Stata v15 won't let me use estat effects to get the design effect of a survey with post stratification calibration of weights? Am I missing some concept that makes a design effect meaningless in such a case? @statstas.bsky.social , @tslumley.bsky.social
Peter EllisNov 12, 2024
Overdue, but from now all statistics procurements involving code from my work will require code be developed collaboratively and handed over on GitHub, peer review via pull requests, Git skills an essential requirement in evaluation of bids. Surprising how many providers this effectively rules out.
Peter EllisNov 11, 2024
In a mid term review workshop for UN's Women Count program I am struck how few people from official statistics I am connected with on Mastodon or Bluesky. More are on LinkedIn. Does anyone know where (if anywhere) the statisticians and economists from national stats offices, UN agencies, etc lurk?
Peter EllisNov 11, 2024
Gelman et al.Nov 11, 2024
Meta-analysis with a single study https://statmodeling.stat.columbia.edu/2024/11/11/meta-analysis-with-a-single-study/
Peter EllisNov 11, 2024
If someone (female, not a historian) posts that Gibbon's Decline and Fall etc 'is the standard because it's the best' and I (male, also not a historian) reply, with link, politely, to effect that modern historians wince on hearing this, am I a mansplaining reply-guy?
Yes
9.1%
No
63.6%
Show results
27.3%
Poll ended at .