iiinova — where your digital self keeps evolving

Andy Garrett @andygarrett.bsky.social · 13/08/2026

The landscape is indeed stunning there. How the beautiful oryx survive is beyond me.

110

Andy Garrett @andygarrett.bsky.social · 06/08/2026

A result of removing friction in the process. Ppl need agents to cope with other people’s agents.

010

Andy Garrett @andygarrett.bsky.social · 03/08/2026

It is failure of some people to negotiate a deal, despite facing oblivion, that gets me every time. It’s the same with Monopoly. Thank you for your attention to this matter.

000

Andy Garrett @andygarrett.bsky.social · 29/07/2026

Clearly people think in multiples of 5 here. Perhaps an indication as to why a global mean temperature increase of 1.5 degrees doesn’t always resonate with the public.

000

Andy Garrett @andygarrett.bsky.social · 23/07/2026

Median survival time is the standard way to report such things. That and the Hazard Ratio. Subjects still alive are censored at that point on the curve (those little sticks) as enrolment happens over many months usually. The survival curve takes censoring into account. These are impressive data.

110

Andy Garrett @andygarrett.bsky.social · 17/07/2026

Time to become a Bayesian. Bring a prior that dominates the data. Impossible to lose. My bet, my prior.

110

Andy Garrett @andygarrett.bsky.social · 16/07/2026

We will be allowed to bet on approaching statistical significance? Or a trend towards?

240

Andy Garrett @andygarrett.bsky.social · 16/07/2026

Pocket calculator would surely have been 100%.

140

Andy Garrett @andygarrett.bsky.social · 15/07/2026

6th explainer from the RSS’s climate change task force.

010

Andy Garrett @andygarrett.bsky.social · 08/07/2026

The Count Binface entry enables a protest vote. A ‘Martin Bell type’ candidate might also have been an interesting one to observe. Reform vs one ANOther is a potential banana skin for Farage.

001

Andy Garrett @andygarrett.bsky.social · 01/07/2026

UK and France being able to maintain the nuclear option is a key deterrent for conventional attack. UK dependent on US here though. EU needs UK and France here.

000

Andy Garrett @andygarrett.bsky.social · 27/06/2026

SAIL etc work on cost recovery basis. There is no token-like costs as far as I am aware. More familiar with administrative data access via digital economy act, but never come across the cost based on number of analyses, once data is accessed via TREs, but will check.

100

Andy Garrett @andygarrett.bsky.social · 27/06/2026

P-hacking is a greater risk when data have already been collected. Pharma has led the way in the clinical trials arena to avoid p-hacking (pre-specification in protocols, SAPs, now estimands) for prospectively collected data. P-hacking using retrospective data is a much wider problem than Pharma.

110

Andy Garrett @andygarrett.bsky.social · 25/06/2026

We have certainly done that with coal. Making decent progress elsewhere.

000

Andy Garrett @andygarrett.bsky.social · 25/06/2026

I think the debate sometimes confuses supply and demand. We need to continue and aggressively reduce our demand for oil and gas and decarbonise. However we also need a reliable supply of oil and gas while we do it, including refining.

100

Andy Garrett @andygarrett.bsky.social · 19/06/2026

TBH it just shows how hard this stuff is to do, and the many traps that are easy to fall into when working with these types observational data - where you have no control over the data generating process. This seems to be “kicking the tyres” on SAIL findings, which is how it should be.

150

Andy Garrett @andygarrett.bsky.social · 17/06/2026

If it has a rugby league team, it is in the North. Warrington in Cheshire is definitely the North.

000

Andy Garrett @andygarrett.bsky.social · 10/06/2026

Latest explainer from our @royalstatsoc.bsky.social climate change task force. This explainer led by @andrewcparnell.bsky.social and Hamish Steptoe. @adrianraftery.bsky.social @dariodomi.bsky.social

000

Andy Garrett @andygarrett.bsky.social · 23/05/2026

rss.org.uk/news-publica... Our latest explainer from the @royalstatsoc.bsky.social climate change task force. A couple more coming out over the next few weeks. @dariodomi.bsky.social @adrianraftery.bsky.social

rss.org.uk

Climate explainer: when and how should climate data be adjusted over time

052

Andy Garrett @andygarrett.bsky.social · 09/04/2026

Trials have certainly become more global, complex and thorough - and investigators find it harder to participate as a result of the increase in visits, procedures and tests. It looks like an interesting article, albeit long. Certainly GCP changed things and brought in needed process documentation.

010

Andy Garrett @andygarrett.bsky.social · 01/04/2026

Thought it was an April 1st joke, but see it is 31st March.

010

Andy Garrett @andygarrett.bsky.social · 14/03/2026

Just checked for him. Last post November 2022 saying that he was seriously considering voting with his paws and leavingTwitter. His latest feed seems to be a video of Pete Hegseth and a dodgy handshake. And Lionel Messi’s dating history, and the bad behaviour of horse rating fans at Cheltenham.

010

Andy Garrett @andygarrett.bsky.social · 14/03/2026

He isn’t showing any signs. Not keen on huskies, but other than that seems well balanced.

091

Andy Garrett @andygarrett.bsky.social · 14/03/2026

My dog still has an X account that I have given up trying to delete. He has no followers and follows no-one since his human family left. I can confirm that he gets the same toxic content.

2278

Andy Garrett @andygarrett.bsky.social · 14/03/2026

www.ametsoc.org/ams/about-am... Statement on Public Availability of Scientific Information and Scientific Evidence on Climate Change. Put simply, lose the data, lose the evidence. Also a nice accessible summary from reputable US scientific bodies on the current state of play.

ametsoc.org

Statement on Public Availability of Scientific Information and Scientific Evidence on Climate Change

The AMS is a global community committed to advancing weather, water, and climate science and service.

010

Andy Garrett @andygarrett.bsky.social · 07/03/2026

Does that matter for set pieces?

000

Reposted by Andy Garrett

Our World in Data @ourworldindata.org · 06/03/2026

Teaching economics or social sciences? Concepts like the unemployment rate or labor force participation can be tricky to explain clearly. Our new Work & Employment page breaks these down with clear explanations and visualizations — all free to download and use in your classes and presentations.

12611

Andy Garrett @andygarrett.bsky.social · 05/03/2026

There is a shift now towards protocol digitisation whereby other things can be generated from that protocol. People spoke about machine readable protocols previously and I honestly don’t know the difference. Essentially it must come down to digitalisation and the associated meta data.

020

Andy Garrett @andygarrett.bsky.social · 05/03/2026

The IT company we worked with at the time didn’t use Word, and that led to limitations (spell checking etc). It was a licensing issue, I recall. And then they switched to Word. I think they ran out of cash eventually. It was when I worked at a different company.

110

Andy Garrett @andygarrett.bsky.social · 05/03/2026

And ppl often cut and paste text and TLF templates anyway from similar protocols, SAPs and libraries. And then sometimes get badly caught out by not changing things that account for the differences. The lack of critical thinking is the issue.

130

Andy Garrett @andygarrett.bsky.social · 05/03/2026

The SAP should be an elaboration of the stats methods etc of the protocol. Produce a decent protocol and the SAP follows on nicely. The LLM generation is an interesting one. I think it depends on what is input. Input a decent protocol and relevant TLF templates, then it could do heavy lifting.

120

Andy Garrett @andygarrett.bsky.social · 05/03/2026

Over 10 years ago we looked at structured content authoring where text was directly linked. Protocol, SAP, CSR in a deterministic way. Amend the protocol and the changes transfer across into corresponding sections of other documents. It didn’t quite deliver, but the concept still feels right.

110

Andy Garrett @andygarrett.bsky.social · 05/03/2026

The time would include all the text and output shells (TLFs), QC, review, iterations and sign-off.

120

Reposted by Andy Garrett

Royal Statistical Society @royalstatsoc.bsky.social · 24/02/2026

You only have until Friday to nominate for RSS Council, and help shape a data‑driven future! @andygarrett.bsky.social found that Council broadened his insight and influence, leading to roles like Climate Change Task Force Co‑chair and even RSS President. Nominate: rss.org.uk/news-publica...!

031

Andy Garrett @andygarrett.bsky.social · 16/01/2026

A provenance casualty - won’t be the first or last to be caught out. It’s a new trap for the incurious.

000

Andy Garrett @andygarrett.bsky.social · 16/01/2026

Short days they may be, this time of year, but you get to see some lovely sun rises when walking the dog. Sun not quite up in Oxwich, Gower with the Bristol Channel in the distance.

020

Andy Garrett @andygarrett.bsky.social · 31/12/2025

Frosty Richmond Park this morning.

030

Andy Garrett @andygarrett.bsky.social · 28/12/2025

Nice to see KK was pleased with the travel clock and aftershave. Different times indeed.

010

Andy Garrett @andygarrett.bsky.social · 28/12/2025

For historical context, played Leeds United in 4th Round. 0-0 and then replay in Leeds. Dispute between the National Coal Board and the National Union of Mineworkers led to miner’s strike. Kick off for replay was 2:30 Wednesday as not allowed floodlights. Leeds won 2:0 @iandoylesport.bsky.social

010

Andy Garrett @andygarrett.bsky.social · 28/12/2025

Back from Liverpool visiting Mum. Found this from the Kevin Keegan Fan Club. Must have been early 1972 based on the games referenced. Football very different then, although arguably Keegan was one of the first superstars. Fan club had just 200 members at this point. Signed by Kevin himself.

210

Reposted by Andy Garrett

Our World in Data @ourworldindata.org · 26/12/2025

Each Friday, we’re sharing highlights from our Top of the Charts, a look back at our most popular charts, articles, & more in 2025. Today we give you our top 5 most-read topic pages this year. 🧵 1. CO₂ and Greenhouse Gas Emissions → ourworldindata.org/co2-and-gree...

511834

Andy Garrett @andygarrett.bsky.social · 20/12/2025

If you had to choose one team in the Premier League for Wirtz to start against, it would be Spurs.

000

Andy Garrett @andygarrett.bsky.social · 16/12/2025

Beautiful sunrise in Kew down by the river this morning - whilst walking the dog. Chiswick Bridge in the distance and the Boat Race finish line.

031

Andy Garrett @andygarrett.bsky.social · 01/12/2025

Not withstanding the long term impact of losing key expertise from some federal institutions. The long-term impact of that is yet to be realised. As is the impact of the data holes. You never find out how important something is until you realise it isn’t there anymore.

020

Reposted by Andy Garrett

Royal Statistical Society @royalstatsoc.bsky.social · 27/11/2025

💡 Thanks to @timandraharkness.bsky.social, Chair of @significancemag.bsky.social editorial board, for discussing this issue’s “Bad Stats” column on Times Radio! 📣 Hear her on Rod Liddle’s show (starts at 40min): www.thetimes.com/radio/show/2... 📖 Read & subscribe: significancemagazine.com

064