Peter Ellis
@freerangestats.info
Director, Statistics for Development Division, at the Pacific Community. Posts are personal views, and most likely to be about data and #rstats, maybe with bits of history, social science and philosophy thrown in. Blog is at https://freerangestats.info/.
Pinned
We are looking for a capacity development expert who is keen on working in the statistics domain, rather than a statistics specialist who has picked up a bit of training experience on the side. Skills in train-the-trainer, adult learning, and advising on organisational change particularly welcome.
Reposted by Peter Ellis
A recent essay claims the poverty line for a family of four should be $140,000. I disagree:
economistwritingeveryday.com/2025/11/26/t...
economistwritingeveryday.com/2025/11/26/t...
The Poverty Line is Not $140,000
A recent essay by Michael W. Green makes a very bold claim that the poverty line should not be where it is currently set — about $31,200 for a family of four — but should be much higher…
economistwritingeveryday.com
November 26, 2025 at 5:19 PM
A recent essay claims the poverty line for a family of four should be $140,000. I disagree:
economistwritingeveryday.com/2025/11/26/t...
economistwritingeveryday.com/2025/11/26/t...
Reposted by Peter Ellis
To get this takes time. We hire software engineers, and they learn with us about epidemiology, NHS data, and TREs. We hire researchers, and they learn with us about GitHub, and what technical product managers do. This is the team that won this award!
November 26, 2025 at 6:01 PM
To get this takes time. We hire software engineers, and they learn with us about epidemiology, NHS data, and TREs. We hire researchers, and they learn with us about GitHub, and what technical product managers do. This is the team that won this award!
Genuine question for historians, putting Trump aside, who would be the 'most openly corrupt president in US history' if we limited it to 2016 and earlier? For comparison purposes.
NEWS --> BBC confirms to me that they did edit a line out of historian @rutgerbregman.com's speech. It called Trump "the most openly corrupt president in US history."
BBC also confirms this was done on the advice of lawyers. So Trump's threats worked.
New piece:
newrepublic.com/article/2036...
BBC also confirms this was done on the advice of lawyers. So Trump's threats worked.
New piece:
newrepublic.com/article/2036...
Trump’s Fury at BBC Gets Unnerving Results with Pro-MAGA Edit Stunner
First, British Broadcasting Corporation execs resigned after Trump complained about a segment. Now the BBC edited out a line from a historian that was critical of Trump. Where does this end?
newrepublic.com
November 25, 2025 at 7:07 PM
Genuine question for historians, putting Trump aside, who would be the 'most openly corrupt president in US history' if we limited it to 2016 and earlier? For comparison purposes.
Reposted by Peter Ellis
Nearly-perfect printed and handwritten text recognition is the most consequential technical contribution to the study of human culture of the last fifteen years, and it's not even close.
It fundamentally changes our (both lay and expert) relationship with the written past.
It fundamentally changes our (both lay and expert) relationship with the written past.
New issue of my newsletter: "The Writing Is on the Wall for Handwriting Recognition" — One of the hardest problems in digital humanities has finally been solved, and it's a good use of AI newsletter.dancohen.org/archive/the-...
The Writing Is on the Wall for Handwriting Recognition
One of the hardest problems in digital humanities has finally been solved
newsletter.dancohen.org
November 25, 2025 at 6:14 PM
Nearly-perfect printed and handwritten text recognition is the most consequential technical contribution to the study of human culture of the last fifteen years, and it's not even close.
It fundamentally changes our (both lay and expert) relationship with the written past.
It fundamentally changes our (both lay and expert) relationship with the written past.
I have been alternating a chapter of @causalinf.bsky.social 's excellent 'Causal Inference the Mixtape' with Enid Blyton's Malory Towers, a notable contrast. Generally works, but a little worried the exhortations to be kind, mature and decent at games may clash with the DDD identification strategy.
November 25, 2025 at 7:47 AM
I have been alternating a chapter of @causalinf.bsky.social 's excellent 'Causal Inference the Mixtape' with Enid Blyton's Malory Towers, a notable contrast. Generally works, but a little worried the exhortations to be kind, mature and decent at games may clash with the DDD identification strategy.
Reposted by Peter Ellis
Some thoughts on empirical distributions of z-scores
statmodeling.stat.columbia.edu/2025/11/24/s...
statmodeling.stat.columbia.edu/2025/11/24/s...
Some thoughts on empirical distributions of z-scores | Statistical Modeling, Causal Inference, and Social Science
statmodeling.stat.columbia.edu
November 24, 2025 at 4:45 PM
Some thoughts on empirical distributions of z-scores
statmodeling.stat.columbia.edu/2025/11/24/s...
statmodeling.stat.columbia.edu/2025/11/24/s...
These are indeed some great photos of birds
November 25, 2025 at 1:43 AM
These are indeed some great photos of birds
Reposted by Peter Ellis
Reposted by Peter Ellis
This, by @giacgrassi.bsky.social, remains one of the simplest and best climate visuals
November 23, 2025 at 9:30 PM
This, by @giacgrassi.bsky.social, remains one of the simplest and best climate visuals
Reposted by Peter Ellis
'I’m not sure which is worse, a total lack of self-awareness or just enough to recognise the problem but insufficient self-discipline to do anything about it.'
It’s too late for Robert Shrimsley, but you can still follow his advice: on.ft.com/4of0ocb
It’s too late for Robert Shrimsley, but you can still follow his advice: on.ft.com/4of0ocb
November 22, 2025 at 10:00 AM
'I’m not sure which is worse, a total lack of self-awareness or just enough to recognise the problem but insufficient self-discipline to do anything about it.'
It’s too late for Robert Shrimsley, but you can still follow his advice: on.ft.com/4of0ocb
It’s too late for Robert Shrimsley, but you can still follow his advice: on.ft.com/4of0ocb
Reposted by Peter Ellis
And it has indeed come to pass.
Trump in a room with a very charismatic person results in like half a day of Trump agreeing with everything that person said, so this has the potential to be very, very funny.
Trump says he’s meeting with “communist” Mamdani at the White House on Friday
November 21, 2025 at 9:15 PM
And it has indeed come to pass.
Reposted by Peter Ellis
Three meta-principles of statistics: the information principle, the methodological attribution problem, and different applications demand different philosophies
statmodeling.stat.columbia.edu/2025/11/20/t...
statmodeling.stat.columbia.edu/2025/11/20/t...
Three meta-principles of statistics: the information principle, the methodological attribution problem, and different applications demand different philosophies | Statistical Modeling, Causal Infere...
statmodeling.stat.columbia.edu
November 20, 2025 at 2:41 PM
Three meta-principles of statistics: the information principle, the methodological attribution problem, and different applications demand different philosophies
statmodeling.stat.columbia.edu/2025/11/20/t...
statmodeling.stat.columbia.edu/2025/11/20/t...
Reposted by Peter Ellis
We are looking for a capacity development expert who is keen on working in the statistics domain, rather than a statistics specialist who has picked up a bit of training experience on the side. Skills in train-the-trainer, adult learning, and advising on organisational change particularly welcome.
November 19, 2025 at 3:13 AM
We are looking for a capacity development expert who is keen on working in the statistics domain, rather than a statistics specialist who has picked up a bit of training experience on the side. Skills in train-the-trainer, adult learning, and advising on organisational change particularly welcome.
The underlying problem here is not the AI bot, but that all these former staff were still on the invite list. Presumably only possible because they were using external email addresses. Most orgs stop this problem for ex-staff by insisting you use your organisational email address.
A group of hospital physicians met for a virtual rounds meeting.
What they didn't realize was that Otter.ai was on the call, too.
They found out when the AI transcriber sent a transcript of the meeting -- patient information and all -- to dozens of current and former staff after the call.
What they didn't realize was that Otter.ai was on the call, too.
They found out when the AI transcriber sent a transcript of the meeting -- patient information and all -- to dozens of current and former staff after the call.
AI bot recorded doctors’ meeting, sent patient info to current and former hospital staff, watchdog says
The transcription tool recorded the meeting on behalf of a physician who no longer worked at the hospital
www.theglobeandmail.com
November 19, 2025 at 7:25 PM
The underlying problem here is not the AI bot, but that all these former staff were still on the invite list. Presumably only possible because they were using external email addresses. Most orgs stop this problem for ex-staff by insisting you use your organisational email address.
Excited we are hiring a "statistical capacity development adviser". This is a one year role in Suva, an investment to improve the quality of all the work we (45 of us) do to promote long term capacity for statistics in Pacific island countries and territories. careers.spc.int/job/statisti...
Statistical Capacity Development Adviser
Suva-based position (Fiji) Attractive expatriate package Join the principal development organisation in the region Description The Pacific Community (SPC) is...
careers.spc.int
November 19, 2025 at 3:11 AM
Excited we are hiring a "statistical capacity development adviser". This is a one year role in Suva, an investment to improve the quality of all the work we (45 of us) do to promote long term capacity for statistics in Pacific island countries and territories. careers.spc.int/job/statisti...
Just found myself writing "I suspect Excel will be appropriate technology for this exercise", somewhat to my surprise.
November 19, 2025 at 2:43 AM
Just found myself writing "I suspect Excel will be appropriate technology for this exercise", somewhat to my surprise.
Reposted by Peter Ellis
On any average day in the last 35 years, about 115,000 people left extreme poverty behind. But based on current trends, progress against extreme poverty will come to a halt.
Important slow news story on @ourworldindata.org
ourworldindata.org/end-progress...
Important slow news story on @ourworldindata.org
ourworldindata.org/end-progress...
November 17, 2025 at 8:02 PM
On any average day in the last 35 years, about 115,000 people left extreme poverty behind. But based on current trends, progress against extreme poverty will come to a halt.
Important slow news story on @ourworldindata.org
ourworldindata.org/end-progress...
Important slow news story on @ourworldindata.org
ourworldindata.org/end-progress...
Reposted by Peter Ellis
This week on the blog: Hoplites! We're taking a crack at explaining the long-running debate over the nature and significance of the ancient Greek heavy infantryman, the hoplite, and the phalanx in which he (mostly) fought.
acoup.blog/2025/11/14/c...
acoup.blog/2025/11/14/c...
Collections: Hoplite Wars, Part I: The Othismos over Othismos
This week (and next) we’re looking at hoplites, the heavy infantry of the ancient Greek poleis in the (early? mid? late?) Archaic and Classical periods, into the Hellenistic. In particular, I…
acoup.blog
November 15, 2025 at 3:09 PM
This week on the blog: Hoplites! We're taking a crack at explaining the long-running debate over the nature and significance of the ancient Greek heavy infantryman, the hoplite, and the phalanx in which he (mostly) fought.
acoup.blog/2025/11/14/c...
acoup.blog/2025/11/14/c...
Reposted by Peter Ellis
Quite interested in how the smoothing of the lines is for statistical accuracy (GAM smooth), not design choices. I find them more readable in general. 📊
Those claiming Dems should retreat on racial justice aren't hard-headed realists, they're pushing against the electoral tide rather than leaning into it. The story of Gen Z isn't about racist backlash or red-pilled young men. It's the most racially progressive generation in American history. 🧵
November 15, 2025 at 3:29 AM
Quite interested in how the smoothing of the lines is for statistical accuracy (GAM smooth), not design choices. I find them more readable in general. 📊
Reposted by Peter Ellis
This means, as a minimum, full computational reproduction of the results from the data. Apart from fraud, this will certainly unearth many honest mistakes because people don't double-check their results if they give them the answers that they hoped for.
November 15, 2025 at 12:35 AM
This means, as a minimum, full computational reproduction of the results from the data. Apart from fraud, this will certainly unearth many honest mistakes because people don't double-check their results if they give them the answers that they hoped for.
Reposted by Peter Ellis
The degree to which this sort of interactivity is not just possible but simple to code in #rstats would have seemed incredible 15 years ago.
The `enable_shiny_hover()` function in mapgl for #rstats lets you use your mouse's position as a Shiny input.
Be judicious with how you use it... but it unlocks so many possibilities for mapping your data.
Example: Census tracts that restyle based on your cursor position
Be judicious with how you use it... but it unlocks so many possibilities for mapping your data.
Example: Census tracts that restyle based on your cursor position
November 12, 2025 at 10:02 PM
The degree to which this sort of interactivity is not just possible but simple to code in #rstats would have seemed incredible 15 years ago.
Reposted by Peter Ellis
All of BLS has been furloughed since October 1st. That means no one was collecting data in Oct. You can’t just walk into a Costco in mid-November and find out what the price of a good was in October. That’s why an October CPI isn’t possible.
October jobs data may never be released, White House says
October jobs data may never be released, White House says
The Bureau of Labor Statistics has been unable to collect price data and employment information during the record-breaking government shutdown.
dlvr.it
November 12, 2025 at 8:25 PM
All of BLS has been furloughed since October 1st. That means no one was collecting data in Oct. You can’t just walk into a Costco in mid-November and find out what the price of a good was in October. That’s why an October CPI isn’t possible.
Reposted by Peter Ellis
File under: "well who could have seen that coming". The Billions to Trillions agenda of using aid to crowd in private finance for development hasn't worked: devpolicy.org/beyond-aid-p...
Beyond aid part 3: rethinking the role of private finance from trillions to realism - Devpolicy Blog from the Development Policy Centre
The ”billions to trillions” narrative has faded and the focus has now shifted to global public finance, says Nicola Nixon et al.
devpolicy.org
November 11, 2025 at 6:05 PM
File under: "well who could have seen that coming". The Billions to Trillions agenda of using aid to crowd in private finance for development hasn't worked: devpolicy.org/beyond-aid-p...
Reposted by Peter Ellis
Really nice to see this being used - a real blast from the past. If anyone is keen to help get this back on CRAN or do other work on it that would be amazing!
@dslc.io welcomes you to week 45 of #TidyTuesday! We're exploring WHO TB Burden Data: Incidence, Mortality, and Population!
📁 https://tidytues.day/2025/2025-11-11
📰 https://samabbott.co.uk/getTBinR/index.html
#RStats #PyData #JuliaLang #DataViz #tidyverse #r4ds
📁 https://tidytues.day/2025/2025-11-11
📰 https://samabbott.co.uk/getTBinR/index.html
#RStats #PyData #JuliaLang #DataViz #tidyverse #r4ds
November 11, 2025 at 10:34 AM
Really nice to see this being used - a real blast from the past. If anyone is keen to help get this back on CRAN or do other work on it that would be amazing!