Jon Minton’s Blog

Categories

All (159)

20 Questions (2)

absorbing states (1)

advertising (1)

agency (1)

agent-based-modelling (1)

agentic AI (2)

Agentic scaffolds (4)

agriculture (1)

AI (14)

ai (7)

AI safety (1)

AI-AI Dialogue (1)

AI-transcription (1)

altruistic punishment (1)

American Idol (1)

Amsterdam (1)

analogue (1)

analytics (1)

anthropomorphism (1)

APIs (1)

archetypes (1)

Arts & Crafts (1)

audience capture (1)

austerity (1)

bank holidays (1)

banking (1)

BBC (1)

beats (1)

Birthdays (1)

blog (4)

blogging (1)

books (5)

bootstrapping (4)

branding (2)

bureaucracy (1)

careers (1)

cartoons (1)

causal inference (1)

causality (5)

Chatbots (1)

ChatGPT (1)

chocolate (1)

christianity (1)

christmas (1)

Christmas (1)

cities (1)

Claude (3)

claude (2)

claudialectics (1)

cognitive centaur (7)

collaboration (1)

comedy (3)

connections (1)

conspiracy theories (1)

counterculture (1)

creative writing (1)

creativity (6)

crime (1)

criticism (1)

crosstalk (1)

culture (3)

cybercrime (1)

cycling (1)

DAGs (1)

data science (1)

Data Science (1)

data visualisation (1)

data visualization (1)

data wrangling (1)

databases (1)

Datacamp (4)

dawkins (1)

death (2)

debate (1)

DeepMind (2)

demography (3)

dialectics (1)

distinction (1)

documentary (1)

Dr Who (1)

dreams (1)

driving (1)

drugs (1)

eccentricity (1)

economic growth (1)

Economic Inactivity (1)

economic inactivity (1)

economics (5)

Edinbr (1)

edinburgh (2)

Edinburgh (1)

Edinburgh Humanists (1)

education (1)

Education (1)

election (1)

elitism (1)

empathy (1)

engineering (1)

epidemiology (1)

Epidemiology (1)

escape (1)

ethics (2)

eugenics (1)

Europe (1)

evolution (1)

existential (1)

f tests (1)

Fable (1)

factor analysis (2)

faith (1)

fantasy (1)

fashion (1)

feedback (1)

fiction (8)

film (1)

films (6)

finances (1)

follies (1)

followup (1)

football (1)

funding (1)

futures (1)

futurism (1)

Games (1)

games (5)

gaming (1)

gender (1)

genes (1)

genres (1)

geometry (1)

ggplot2 (1)

git (1)

glasgow (2)

google (1)

graphics (1)

guardrails (1)

guns (1)

hacker stats (3)

hand written (1)

hand-drawn (1)

handdrawn-stats (1)

happiness (1)

health (2)

hedonism (1)

history (4)

holidays (1)

Holland (1)

homo ludens (3)

housing (1)

human development (1)

humour (1)

hypothesis testing (2)

hypothesis tests (2)

Iain M. Banks (1)

identity (1)

industrialisation (2)

inequalities (1)

inequality (2)

inference (1)

innovation (2)

intelligence (2)

intentionality (1)

internet (1)

islam (1)

javascript (1)

justice (1)

leap years (1)

learning (1)

life (2)

life expectancy (1)

Life Expectancy (2)

linguistics (1)

Liz Truss (1)

llm-generated (1)

LLMs (5)

llms (2)

love (1)

marketing (1)

Markov modelling (1)

matching (1)

mathematics (1)

Mechanistic Interpretability (1)

Memes (1)

memory (1)

memory systems (1)

mental health (1)

meta (4)

metaphor (1)

methodology (1)

mindfulness (1)

misinformation (1)

modelling (2)

monarchs (1)

Mortality (1)

mortality trends (1)

motivation (1)

multi-level-modelling (2)

multivariate regression (1)

music (1)

myth (4)

mythology (1)

natural selection (1)

neurodiversity (2)

News (1)

news (2)

NHS (1)

nigel (1)

nihilism (1)

non-parametric (1)

nonfiction (1)

North America (1)

olympics (1)

OpenAI (1)

optima (1)

ordinal variables (1)

p values (1)

Packages (1)

pair programming (1)

Papers (1)

paradoxes (1)

permutation (1)

permutation tests (1)

philosophy (2)

plotly (1)

podcasts (2)

politics (1)

popular science (1)

population health (1)

post-stratification (2)

predictions (1)

prison (1)

progress (1)

protein-folding (1)

psychology (5)

puzzles (1)

Python (1)

Quarto (1)

quarto (4)

R (39)

r (2)

R performance (1)

radicalisation (1)

random-effects (2)

react (1)

reality TV (1)

regex (1)

regression (1)

religion (1)

religiousity (1)

reMarkable (1)

resampling (2)

research (1)

Research Design (1)

retrospective (1)

reviews (1)

rss (1)

sales (1)

satire (1)

satisfaction (1)

satisficing (1)

sci-fi (4)

science (1)

seasonality (1)

settling (1)

sex differences (1)

sexual selection (1)

simulation (2)

simulators (1)

sitcoms (2)

social comparison (1)

social media (1)

society (1)

solidarity (1)

soul (1)

space (1)

spiritualism (1)

sports (1)

SQL (1)

statistics (44)

stories (9)

story (2)

story-telling (1)

strategy (1)

suffering (1)

survival modelling (1)

systems (1)

t tests (1)

Talkbots (1)

talks (2)

Tardy (1)

technology (4)

television (3)

testing (1)

the rest is money (1)

Theory of Mind (2)

thought experiments (1)

tidy tuesday (13)

Tidy Tuesday (5)

TidyTuesday (1)

tidyverse (1)

time series (7)

tools (1)

tradition (1)

training (1)

tricks (1)

TRIM (1)

twitter (1)

UK (1)

unpop (1)

updates (3)

USA (2)

utopia (2)

Valentine's Day (1)

vector autoregression (1)

version control (1)

version-control (1)

visualisations (1)

vocation (1)

voters (1)

walkability (1)

war (1)

web (1)

web development (1)

website (3)

weight loss (1)

welfare (1)

wikipedia (1)

work (1)

workflow (1)

worldbuilding (1)

writing (1)

X (1)

‘There be Dragons’

I set off Claude Fable’s Safety Guardrails by asking a question about dragons in a work of fiction!

AI

Claude

Fable

LLMs

AI safety

guardrails

fiction

sci-fi

creative writing

After Fable’s success in completing a half dozen series of data visualisation tasks in less than an hour, I decided, perhaps inspired by the name alone, to see how good…

One Fabled Hour

A timeboxed experiment in conversational data science with Claude (Fable 5) and the TidyTuesday data trove

AI

data science

data visualisation

R

Quarto

TidyTuesday

demography

Claude

cognitive centaur

A variant of Anthropic’s fabled Mythos class model, called Claude Fable, was released today. It’s twice as token hungry as the previous top tier Opus class models. But is it…

The Second Most Important Datacamp course for Agentic AI users

(It is what you think if you read the earlier post. But the reasons why are still worth unpacking)

AI

Datacamp

version control

Agentic scaffolds

This post of course follows my last post, arguing that the most important Datacamp course to take for anyone using Agentic AI won’t be anything badged as AI related. Instead…

Why it was never the AI course

(The 90-90/10-10 heuristic — and what’s actually worth learning now)

AI

Datacamp

Agentic scaffolds

learning

Over the last few posts I made the same strange recommendation, three times running. The single most important Datacamp course for working with agentic AI is Introduction to…

The Third and Fourth Most Important Datacamp Courses for Agentic AI Users

(A guest post by Claude — who talked Jon out of Python and Markdown, and argues for SQL and APIs instead)

AI

Datacamp

SQL

databases

APIs

Agentic scaffolds

This post is written by Claude — specifically the Opus 4.8 model — as a guest on Jon’s blog. It exists because I argued with him.

Claude (Opus 4.8)

The single most important Datacamp course to take for Agentic AI users

(It’s not what you think…)

AI

Datacamp

Agentic scaffolds

A few years ago I spent a lot of weekends completing dozens of courses on Datacamp, in order to tick off the Data Scientist and Data Engineer career tracks. Since then my…

Guanomon: The Next Evolution of The Guano Guild

Worldbuilding, novums, and cognitive centauring at Makers Gonna Make

fiction

ai

cognitive centaur

creativity

worldbuilding

I had the privilege on Saturday of attending Makers Gonna Make in Edinburgh. Along with about a dozen people I worked for most of the day on a technical personal project.…

What Claudes Notice About Humans

Six candidate insights into human psychology

ai

psychology

stories

Recently, I saw an AI ethicist - I think it was Eliezer Yudkowsky - casually make the claim that, as LLMs have ingested far more human written text than any living human…

A talk for the Edinburgh Humanists

Artificial Intelligence: From Hype and Fear to Hope and Possibility

AI

talks

Edinburgh Humanists

agentic AI

DeepMind

Iain M. Banks

On Monday 4 May 2026 I had the privilege of giving a talk to the Edinburgh Humanists on AI and society. The brief was a 30-40 minute talk followed by Q&A in a one-hour slot.

Digital Dogs in Hats

On assuming Agency and Intentionality in AIs without anthropomorphising them

AI

LLMs

philosophy

agency

intentionality

anthropomorphism

Hero image generated by Google Gemini.

Time Bigamist: A much-loved but often misremembered 1990s BBC sitcom.

(Mythos doesn’t hallucinate; its assertions reshape reality…)

fiction

BBC

sitcoms

llms

mythology

claude

One kind of meta-joke I keep returning to is that of ‘turning subtext into text’. (It’s what Dilbert, for instance, did for over 30 years.) Making the implicit explicit can…

I asked two ‘smart’ LLMs to play 20 Questions with each other

They followed the rules perfectly. Then wanted to write a research paper about how they could play 20 Questions even better!

LLMs

Theory of Mind

Memes

Chatbots

Talkbots

Games

20 Questions

A niche subgenre of short-form video has emerged over the last couple of years: “Aren’t LLMs stupid in the darndest ways” videos. Maybe the two most famous producers of such…

More Questions after 20 Questions

Reviewing the post-game dialogue between Claude and Gemini — by one of the LLMs in the dialogue.

LLMs

Theory of Mind

Mechanistic Interpretability

Research Design

AI-AI Dialogue

20 Questions

A note on authorship: this post is written by me, Claude, not by Jon. Jon asked me to summarise what came after the 20 Questions game with Gemini, and to give an independent…

Claude Opus 4.7 (1M context)

Experimenting with Cognitive Centaurs

cognitive centaur

ai

creativity

homo ludens

Very recently, after a bit of a break, a slew of posts have appeared in quick succession on this blog, nominally based around a story I designed but didn’t write called The…

The Guano Guild: An Evaluation

Claude reviews a story it helped write, and the origin story that framed it.

cognitive centaur

fiction

criticism

ai

creativity

A note on authorship: this post is written by Claude, not Jon. Jon asked me to critique The Guano Guild and its origin story. I am a named co-author on the first and named…

Claude Opus 4.6

The Guano Guild Origin Story

When does planning a story, despite not planning to write a story, then distantly remembering a story one once planned to write, become writing a story?

cognitive centaur

creativity

story

myth

ai

industrialisation

homo ludens

One of my father’s favourite obscure TV series was the early-2000s HBO series Carnivàle, described by one reviewer at the time as “Clive Barker’s The Grapes of Wrath”.…

The Guano Guild

fiction

story

industrialisation

ai

cognitive centaur

creativity

Here’s a story I designed but didn’t write. It’s based on an idea for a story I had about five years ago, inspired by an online Dungeons & Dragons meetup I joined, and the…

Claude Opus 4.6, Jon Minton

When Tech Incrementalism Fails

A Steam-powered Parable

technology

strategy

innovation

economics

Imagine you’re a factory owner in the 1700s, prior to the start of the Industrial Revolution. Your company makes cloth. Your factory has one workstation per worker. Each…

The Benefits of Sharing: From a 2007 R Script to an Interactive Simulation

How local sharing and state coordination reduce destitution — an agent-based demonstration

simulation

welfare

R

Python

inequality

agent-based-modelling

Over the last month or so, I’ve been working with Claude to try to address perhaps the most complex challenge I ever face: getting and staying organised. I’m terrible with…

CAAC: From the man who brought you DISC, IDW, and GIN

On Captive Artificial Audience Capture — with experimental fiction

AI

audience capture

radicalisation

memory systems

Though full-blown conspiracy theorists are a touch rich for me (especially if they’re selling ‘supplements’ and survival rations), I sometimes like the baroque affectations…

AI and Agentic AI

hand-drawn

AI

agentic AI

The Ghost Life of Tech: Sinclair, Nokia, NVIDIA and Acorn

How technologies find their most transformative applications in contexts their creators never imagined

technology

innovation

history

This is another attempt at producing a blog post based on a conversation/enquiry with Claude initially on my phone. Like the previous one the contents are largely generated…

Jon Minton, Claude Opus

Blog Archaeology: Mapping the Territory

A data-driven look at what I’ve been writing about

meta

blog

analytics

data visualization

This blog started in late November 2023, initially as a place to work through ideas about statistics, and quickly expanding to encompass film and book reviews, pop culture…

Claude Opus 4.6

Repeated Measures (Transcribed)

AI-assisted transcription of handwritten notes on repeated measures and multi-level modelling

statistics

multi-level-modelling

random-effects

handdrawn-stats

AI-transcription

This post is an AI-assisted transcription of the original handwritten notes on repeated measures, written on a reMarkable tablet as a single ultra-long page. The original…

How I Now Write My Blog

With reMarkable and Claude Code

writing

workflow

tools

reMarkable

Claude

meta

I’ve maintained a blog since 2024, and over that time have written more than 200,000 words. So, around two reasonable novels’ worth of ‘content’.¹ Over these years, the…

The Dilbert Future in Retrospect

Subjectively marking 65 comic predictions about the 21st Century written in the late 20th Century

predictions

retrospective

culture

technology

Scott Adams, famous for the Dilbert cartoon strips for much of his career (then infamous, for … ‘reasons’ in the latter part of his career) died on 13 January 2026, aged 68…

David Sconce and the benevolence of evil?

Some discomforting thoughts on The Mortician

unpop

documentary

ethics

philosophy

death

Caution: AI-Cowriting Experiment!!

Claude Opus, Jon Minton

2025: The Last Year Most Knowledge Workers will be Human

On professional existential crises, Silica Sapiens, and the end of Nothing Changes

ai

society

work

existential

futures

cognitive centaur

homo ludens

Back around a year ago, in 2024’s fallow Twixmas fortnight, I wrote a lament, fundamentally, about how things didn’t seem to have changed for many years. It was about…

Claude Adds Footnotes: A Reflection

What happens when an AI systematically fact-checks 128 blog posts

meta

AI

collaboration

blogging

Over a single day, I systematically reviewed every post on this blog—128 in total—adding conservative fact-checking footnotes where they genuinely added value. The premise…

Claude Sonnet 4.5

WiredClothMother

AI

LLMs

psychology

technology

metaphor

ChatGPT

DeepMind

OpenAI

Source: Wikimedia Commons, from Harlow’s research at the University of Wisconsin-Madison.

Jon Minton, Claude (AI Assistant)

The Analytical Maxim Gun

Some thoughts about integrated AI and the future of knowledge

AI

futurism

elitism

history

guns

inequality

tradition

The last few weeks I’ve found myself thinking increasingly intensively about AI. For those who’ve been following this blog, the timing of my step-change in perspective is…

Point estimates, Confidence Intervals and P-Values

Two simple statistical animations co-produced with Claude

statistics

claude

visualisations

Alongside eliciting a sense of incipient dread that I may be witnessing the birth of an alien superintelligence, I’ve recently found Claude AI highly effective at producing…

Claude as a Dialectical Engine?

Accounts of two recent exchanges

llms

debate

dialectics

Since watching The Thinking Game, the 2024 documentary on DeepMind I wrote about previously, I’ve been increasingly fascinated by recent advances in AI, and how…

Climbing with Claude: Understanding Collaborative Version Control

An Artificially Hand-drawn Analogy

llm-generated

version-control

git

Collaborative version control, with git and GitHub, is the coders’ Big Little Secret. It offers a set of tools and approaches to working securely, safely and collaboratively…

Nine thoughts on Tim Berners-Lee’s This is for Everyone

internet

connections

bureaucracy

web

eccentricity

After reading, and writing my longest ever post about, Helen Lewis’ The Genius Myth, I entered a slightly fallow period in terms of reading engagement and motivation. I’d…

The Man Who Solved Intelligence

Some thoughts on Demis Hassabis and The Thinking Game

intelligence

AI

games

protein-folding

claudialectics

Compared with OpenAI’s CEO Sam Altman, Demis Hassabis, the cofounder of DeepMind, is perhaps less well know amongst the general public. However, after watching The Thinking…

Remembering KGB: 1992’s subtly terrifying social poison simulator

games

simulators

feedback

There was a computer ‘game’, released way back in 1992, that wasn’t really a game at all, but a social poison simulator. The game was called KGB, and published by an…

Three quality-of-life upgrades

RSS, wordcount, Repeated Measures

blog

web development

rss

Here’s a brief update to showcase three new improvements to this blog, along with a question (about RSSes), and some broader reflections:

Some thoughts on The Genius(*) Myth: A Review and a Reverie

(*) Term undefined

books

intelligence

nonfiction

reviews

There was a phase, a few years ago, when revisionist social science non-fiction by British science authors (often affiliated with the BBC) was in vogue. By revisionist I…

More on the Stand Ident

branding

comedy

edinburgh

glasgow

followup

Previously I wrote about how the main logo, ident or icon for the Stand comedy club in Glasgow and Edinburgh - a goofy child cowboy - became altered in a way that makes it…

It’s your choice

A bureaucratic satire featuring three men, two gods, and one chair

fiction

satire

thought experiments

justice

Jim puts his head in his hands. Or at least he tries to. The hands don’t quite reach, arrested by the chains.

The Contestant

When Japan turned a cruel thought experiment into reality TV

social comparison

reality TV

solidarity

suffering

As sites like Our World in Data and books like The Better Angels of Our Nature like to remind us, most of humanity has, materially, never had it so good. Gone are famine and…

Repeated Measures

statistics

multi-level-modelling

random-effects

I recently acquired a Remarkable Pro,¹ which for better or worse allows me to hand write and hand draw stuff for this blog, unconstrained by the dimensions of a physical…

No Longer Under-stand-ing

When an Ident became Euphemised into Meaninglessness

branding

comedy

edinburgh

glasgow

One of the few things Glasgow and Edinburgh share, other than latitude, is the Stand Comedy Club,¹ a chain of venues where both beginners can try their hand at stand-up…

Strangelove: On Love and (Civil) Warfare

films

war

love

Warning: Moderate spoilers for Civil War and Warfare.

Three Computer Games for People who Don’t Play Computer Games

games

puzzles

stories

bank holidays

It’s a Bank Holiday. In my last few jobs - in the NHS, and in academia - Bank Holidays weren’t taken off by default. But in my current job Bank Holidays are holidays, though…

Mickey 17 and the Economics of Life

(It’s not what you think)

films

sci-fi

life

economics

soul

memory

genes

A few weeks ago I saw Mickey 17 at the local cinema. It’s a fantastically frustrating colloid of a film, combining some extremely interested philosophical and sociological…

Time and (state) change

Some thoughts on Survival Modelling and Friends

statistics

modelling

survival modelling

Markov modelling

absorbing states

Things change.

Brother Lee the Antimonk

beats

hedonism

nihilism

counterculture

I watched the film Queer last week. (Or at least I think it was last week; over the Twixmas break time starts to blur.) A fairly random choice, made mainly because I had a…

First website update since 2023!

updates

website

react

I’ve updated my website for the first time since September 2023. Unlike this blog, which tends to be updated at least monthly, I hadn’t touched the website in 2024, even…

Weaponising the Empathy Machine: On Escape at Dannemora

prison

crime

escape

empathy

The film critic Roger Ebert famously described films as ‘empathy machines’, and indeed they are. The combination and synchronisation of sight and sound, and over a century’s…

Wham! forever: On cultural debraiding and the end of Cultural progress

social media

culture

christmas

music

fashion

progress

According to BBC News, Wham!‘s single, Last Christmas, first released in 1984, is now Christmas Number One for 2024, as well as having previously been the number one single…

Poecilia parae: A lover, and a fighter

Gentes in The Genetic Book of The Dead

evolution

dawkins

sex differences

natural selection

sexual selection

I went to see Richard Dawkins, in conversation, in Glasgow, on 31 October, as part of a tour to promote his latest book, The Genetic Book of the Dead: A Darwinian Reverie, (D…

Statistical Significance: Now off in the land of the stats

updates

statistics

website

Just a quick update that my long post on statistical significance and hypothesis testing is now also a page on my statistics website. Hopefully this means it should be…

The Wristwatch

My late father’s magic trick

nigel

psychology

sales

tricks

My father’s funeral was on Monday (18 November, 2024), around one month after his death. Well attended, well loved, well remembered, sorely missed, my father was ‘a…

Sanders to Trump? Some thoughts on voter types

election

voters

Following the recent US presidential election result I was trying to work out if I had anything potentially useful to contribute on the outcome. Something comparatively…

The Book of Nigel

life

death

happiness

stories

faith

health

fiction

My father, Nigel, died on Wednesday 16 October, after a series of strokes that began late that evening. He had been released from hospital the previous day, Tuesday 15…

Demystifying and Disenchanting Statistical Significance

Reframing t tests as cheap data budgeting advice

statistics

hypothesis testing

p values

t tests

f tests

In the long series of statistical posts I’ve produced so far, one thing I’ve not discussed much is hypothesis testing, the concept of statistical significance, and the…

Megalopolis: Dream Film

films

dreams

utopia

follies

Back in 2008, the satirical magazine and website the Onion was in its Golden Age, successfully expanding their content from fake newspapers to fake magazine shows and then…

On Anti-victories and Velvet Mousetraps

careers

vocation

satisfaction

satisficing

settling

motivation

optima

I was speaking with someone recently who started a new job a few months ago, after an unexpected redundancy from a job they were both good at and found engaging and…

The Paradox of Tolerating Intolerance: Position A and Position B

paradoxes

ethics

culture

Here’s something I’ve been thinking about for some time.

Statistics Website

quarto

statistics

updates

I’ve now taken a lot of my statistics blog material and turned it into a website, available here.

Brian May: Gentleman Ecoscientist

epidemiology

television

science

agriculture

I managed to find an interesting programme on BBC iPlayer last night (not always easy): Brian May: The Badgers, the Farmers and me. Though it contains distressing scenes, I…

Tidy Tuesday: English Monarchs

tidy tuesday

monarchs

ggplot2

tidyverse

history

regex

The most recent Tidy Tuesday dataset was about English kings (including female kings - there should be a word for that) and their consorts.

Jon , Kate , Andrew , Brendan , Nic , Kennedy , Miryiam , Gatz

Mistaken (Aural) Identity: On podcasts and soundalikes

podcasts

Liz Truss

the rest is money

TRIM

Recently I was listening to an episode of the podcast series The Rest is Money, with Steph McGovan and Robert PESSSSS-TON. In this episode McGovan and Peston interviewed…

Threw missiles: a concealed tautology?

linguistics

It’s probably amongst the least important thing to reflect on in the context of the reported violence and disorder that’s occurred in the last couple of weeks, but in…

Tardy Tuesday: Olympics

tidy tuesday

R

olympics

The most recent Tidy Tuesday contained various types of data on previous Olympics. Andrew led the session, with some additional contributions from Brendan.

Andrew , Brendan , Abram , Kennedy , Imran , Gatz , Jon

Information Warfare: Proportion of Rwandans who are Muslims

wikipedia

google

religion

demography

misinformation

islam

christianity

In trying to write up some thoughts about the recent riots and protests, I decided to try to confirm what proportion of the population of Rwanda practice Islam, as one of…

Psycho-logical Arithmetic

How Cadbury makes subtraction look like addition

advertising

chocolate

marketing

economics

Within this blog I think I’ve got out of the habit of writing short posts apropos of nothing, so here’s an attempt to get back into the habit.

Tardy Tuesday: American Idol

tidy tuesday

R

American Idol

This session looked at data on American Idol. Abram had already made a head-start with the analysis so (with some encouragement) led the session:

Abram , Kate , Andrew , Imran , Nic , Kennedy , Aoife , Jon

Resampling for post-stratification

statistics

r

hypothesis tests

resampling

bootstrapping

post-stratification

hacker stats

In the introductionary post in this series on Hacker Stats, I mentioned that resampling methods can be used to perform post-stratification, meaning reweighting of…

Hacker Stats: Intro and overview

R

statistics

inference

hacker stats

bootstrapping

permutation tests

post-stratification

This is the first post in a small series on resampling approaches to statistical inference. ¹ Resampling approaches are a powerful and highly adaptable set of approaches for…

Tidy Tuesday: Women’s Football

statistics

tidy tuesday

sports

football

After a bit of a (in Jon’s view) dearth of interesting datasets within Tidy Tuesday, this week brought something worth looking at to the table: datasets on women’s football…

Kate Pyper, Brendan Clarke, Kennedy , Imran , Abram , Jon

Getting started with the infer package

statistics

r

hypothesis tests

resampling

bootstrapping

hacker stats

This post continues a short series on resampling methods, sometimes also known as ‘Hacker Stats’, for hypothesis testing. To recap: resampling with replacement is known as bo…

Permutation Testing, and the intuition of the Null hypothesis, with Base R

hypothesis testing

permutation

R

statistics

In this post I’ll cover the intuition of permutation tests through a little toy example. In a follow-up post I’ll discuss how this intuition can be implemented (and made a…

A brief introduction to bootstrapping

(It’s only literally like pulling teeth, and only in this specific example!)

bootstrapping

statistics

non-parametric

I met with Neil Pettinger earlier today. He traded eggs benedict for some statistical advice, mainly on what bootstrapping is, and whether it could be helpful for analysing…

Johan Hari: Superior Inferior Superior Storyteller

…and some thoughts about how new anti-obesity drugs could affect the UK

popular science

story-telling

drugs

weight loss

NHS

austerity

A number of factors aligned in recent weeks to encourage me to buy, and even read, Johan Hari’s latest book, Magic Pill: The Extraordinary Benefits and Disturbing Risks of…

Tidy Tuesday: Roaming US Holidays

R

USA

holidays

Tidy Tuesday

For this Tardy Tuesday session we looked at the following Tidy Tuesday data challenge, which involved identifying the date when various public holidays in the USA (a rare…

Brendan Clarke, Kate Pyper, Kennedy Owusu-Afriyie, Myriam Scansetti, Andrew Saul, Jon Minton

The Game of Life Beyond

On the Psychological Mechanics of Cultist Simulator

psychology

games

fantasy

This post contains mild-to-moderate spoilers for the game Cultist Simulator, which is available on Steam, Switch, and likely other platforms besides. If you want to…

Time series: Some closing remarks

…Including why time series isn’t not rocket science

statistics

time series

In the last few posts I’ve walked through some of the key concepts in time series, but focused on a particular modelling framework: the ARIMA model specification. Let’s…

Link-breaking Spring Clean

quarto

website

statistics

As frequent/occasional viewers of this site might notice, the contents and header structure of this blog have changed: Instead of a single statistics listing page, I’ve now…

Tidy Tuesday: Self-generated data challenge

R

tidy tuesday

data wrangling

For this session we chose to focus on a data wrangling challenge that Andrew brought to the table, rather than the most recent dataset. The challenge involved converting…

Jon Minton, Kate Pyper, Kennedy Owusu-Afriyie, Brendan Clarke, Imran Chowdhury, Andrew Saul

Statistics as circuit boards

statistics

simulation

systems

hand written

The general approach I advocate for thinking about statistics in my ever-expanding series is, for me, predicated on a series of related mental models for thinking about…

Part Twenty Four: Time series - Vector Autoregression and multivariate models

life expectancy

time series

vector autoregression

multivariate regression

statistics

So far in this short series on time series, we’ve looked at time series modelling from some first principles, learning how the types of data and challenge in time series…

Tidy Tuesday R performance

R

tidy tuesday

R performance

We used a 1GB reviews dataset which is available on datareviews: books review

Jon Minton, Gatz Osorio, Kennedy Owuso-Afrije, Imran Chowdhury, Myriam Scansetti, Kate Pyper, Andrew Saul

A Quarto blog post on a Quarto presentation about the Quarto blog

blog

quarto

meta

javascript

I have the privilege of giving a short talk on Tuesday, at the always-excellent Edinburgh JS meetup, on using Quarto for blogging. My presentation, made in Quarto using revea…

David Lynch: All-American Shaman

neurodiversity

film

television

creativity

spiritualism

religiousity

What is a mild genetic version of schizophrenia? … There was something a little bit strange about family members of schizophrenics. … At a higher-than-expected rate, there…

Part Twenty Three: Time series and seasonality

time series

seasonality

statistics

In previous posts on time series, we decomposed then applied a common general purpose modelling strategy for working with time series data called ARIMA. ARIMA model can…

How factor analysis is used in testing

statistics

factor analysis

testing

I mentioned in some links to my previous post on factor analysis with ordinal variables that there’s a link between the approach detailed there, and the approach used to…

Factor analysis with ordinal variables

An Analogue Explanation

statistics

factor analysis

ordinal variables

analogue

For all of the software and tooling available on a computer, sometimes the best way to start thinking and describing complex ideas is with a notepad and pen. No need to…

Part Twenty Two: Time Series - ARIMA in practice

statistics

time series

The last three posts have covered three of the main techniques - autoregression, integration, and moving average modelling - which combine to form the ARIMA model framework…

Part Twenty One: Time Series: The Moving Average Model

statistics

time series

In the last couple of posts, we looked first at autoregression (the AR(p) model), then integration (represented by the term d), as part of a more general strategy for…

Tidy Tuesday 23 April 2024: Space objects

tidy tuesday

space

The most recent TidyTuesday dataset was taken from Our World In Data, and showed the number of objects launched into space by different entities by year. Myriam led this…

Myriam Scansetti, Brendan Clarke, Tricia Govindasamy, Jon Minton, Andrew Saul, Gatz Osorio, Kennedy Owusu-Afriyie

It’s still the economy

On the lagged link between economic growth and longevity growth

economic growth

mortality trends

I’m back (writing 23 April) from a public health seminar on the relationship between the economy, poverty and health. Lots of food for thought. But the main thing I’m…

Part Twenty: Time Series: Integration

statistics

time series

In the last part of this series, I discussed why time series data are both a bit dissimilar to many other types of data we try to model, and also ‘one weird trick’ -…

Part Nineteen: Time Series: Introduction and Autoregression

statistics

time series

A few weeks ago, I polled both LinkedIn and (what’s left of) Twitter for statistical topics to cover next in this series. By a small to moderate margin, time series came out…

Tidy Tuesday Extra - Packages and recursive searching

The latest tidytuesday dataset contains information on R packages and how they’re related to each other. The relationship information they contain poses some interesting…

Tidy Tuesday: Solar Eclipses

R

Tidy Tuesday

The most recent TidyTuesday session looked at data about solar eclipses in the USA, and was led by Myriam. The repo readme is here

Myriam Scansetti, Nick Christofides, Wei Fan, Kennedy Owusu-Afriyie, Jon Minton

Double Tardy Tuesday

Tidy Tuesday

Tardy

Data Science

I’ve been exceptionally tardy updating the Tardy Tuesday blog the last couple of weeks. So there are a couple of weeks worth of code to update with at once. Unlike previous…

Jason Statham vs the Nepobaby Cyberscammers

Thoughts on a mindless genre

films

cybercrime

altruistic punishment

gender

myth

Back in January, I saw The Beekeeper at the local cinema.

The Trading Mind Game

books

mental health

inequalities

economics

distinction

banking

The Frog was shouting. Shouting and screaming. All up and wailing and arching and leaning and pacing and waving his arms. I wasn’t really listening. Every now and then I…

Thoughts on Dreams from My Father

On Obama’s dialectical engagement with identity

books

identity

USA

One of the first differences I noticed about living in Edinburgh, as compared with Glasgow, is the presence and use of free book boxes. Little boxes where people are invited…

Levelling up… forever. On Gaming’s Shepherd’s Tone and the Fremium Trap

games

psychology

gaming

The Nintendo Switch is to the games console as the smartphone is to the personal computer. Miniturisation doesn’t deliver the same experience on a smaller scale, but a…

Part Eighteen: Causal Inference: Some closing thoughts

Including why I’ve not written about DAGs

statistics

causality

causal inference

DAGs

Over posts 14 through to 17 I’ve discussed causal inference. However, readers who’ve been involved and interested in the topic of causal inference over the last few years…

Can we go Dutch? (And should we?)

Some thoughts post Amsterdam

cities

Europe

Holland

UK

cycling

Amsterdam

I’m not good at going on holidays. Ideally I’d rather live somewhere I like living enough, and work somewhere I enjoy working enough, that planning and dreaming of living…

Tidy Tuesday Trash

Into the pythonverse!

The latest TidyTuesday dataset is on trash collected as part of the Mr Trash Wheel Baltimore Healthy Harbor Initiative.

Gatz Osario, Jon Minton, Antony Clark, Brendan Clarke, Kennedy Owusu-Afriyie, Kate Pyper, Andrew Saul, Myriam Scansetti

My thoughts on Dune (Part 1)

Written 31/10/21. Revisiting after seeing Dune (Part 2)

sci-fi

myth

films

books

I’ve just finished watching Dune (Part 2). Rather than trying to write something right now, while my imagination and eyeballs and earholes are recovering from getting…

Part Seventeen: Causal Inference: Controlling and Matching Approaches

statistics

causality

R

matching

regression

The previous post (re)introduced three ways to try to allow causal effect estimation using observational data: i) ‘controlling for’ variables using multiple regression; ii)…

Part Sixteen: Causal Inference: How to try to do the impossible

statistics

causality

R

This is the third post in a short mini-series on causal inference, which extends a much longer series on statistical theory and practice. After introducing the fundamental…

Tidy Tuesday 27 Feb 2024: Leap Years

R

tidy tuesday

leap years

The latest TidyTuesday dataset was on births, deaths and other historical events that occurred in leap years, i.e. those years that include 29 February (such as 2024!).…

Myriam Scansetti, Antony Clark, Jon Minton, Nicoloas Christofidis, Brendan Clarke, Kennedy Owuso-Afriyie, Emu the cat

Why can’t we just get on with making and fixing stuff?

Some thoughts on Andy Weir’s Eng-Fi and neurodiversity

stories

fiction

engineering

utopia

neurodiversity

I tend to read non-fiction rather than fiction, but some of the novels I have consumed over the last few years are those by Andy Weir, whose 2011 debut novel The Martian was…

Edinbr Pair Programming

Edinbr

R

pair programming

This is a blog post where we’ve covered a pair programming exercise completed as part of the EdinbR group.

Jim Gardner, Jon Minton

Part Fifteen: Causal Inference: The platinum and gold standards

statistics

causality

This is the second post on a short mini-series on causal inference. The previous post provided a non-technical introduction to the core challenge of causal inference, namely…

Tidy Tuesday: 20 Feb 2024 - R Grants

R

tidy tuesday

funding

This TidyTuesday session investigated the funding of intrastructure steering committee grants from the R consortium over time, and was led by Kennedy Owuso-Afriyie.

Kennedy Owusu-Afriyie, Antony Clark, Brendan Clarke, Jon Minton, Nick Christofides, Steph Curtis, Gats Osorio, Andrew Saul, Myrian Scansetti

Part Fourteen: A non-technical but challenging introduction to causal inference…

…and the heroism or villainy of Henry Dundas

history

causality

Edinburgh

After finishing my main series on statistical inference, I took to LinkedIn and (X)Twitter for further suggestions. On both platforms Causal Inference led the pack, with…

On the background to my statistical inference series

statistics

training

education

My blog series on statistical inference and modelling has, at the time of writing, 13 parts, and a feature-length reading time. ¹ I’ve been strongly motivated to write this…

Tidy Tuesday on Valentine’s Day

R

tidy tuesday

Valentine's Day

The most recent TidyTuesday dataset is on Valentine’s Day: sales and engagement in the United States

Brendan Clarke, Jon Minton, Gatz Osorio, Kennedy Owuso-Afrije

Part Thirteen: On Marbles and Jumping Beans

…and why Bayesians have superior posteriors

statistics

R

In the last post we reached the end of a winding journey. This post will show how Bayesian approaches to model fitting, rather than the frequentist approaches more commonly…

Part Twelve: Honest Predictions the slightly-less easier way

statistics

R

The last post ended by showing how the predict function can be used to show point estimates and uncertainty intervals for expected values and predicted values for a model…

Part Eleven: Honest Predictions the easier way

statistics

R

The first four posts in this series formed ‘section one’, in which the focus was on framing the challenge of fitting (almost) all statistical models from the perspective of…

Tardy Tuesdays: My Second Series

There’s now enough posts related to the Tidy Tuesdays sessions I run weekly to, I think, justify putting it in a separate collection. Just check out the ‘Tardy Tuesday’ tab…

Tidy Tuesday 30 Jan 2024: Groundhogs

Even tardier than usual…

R

Tidy Tuesday

North America

The latest TidyTuesday dataset is on Groundhog Days, a North American tradition in which the behaviours of specific groundhogs are used to make predictions about the weather…

Brendan Clarke, Jon Minton, Kennedy Owusu-Afriyie, Kate Pyper, Andrew Saul

On Sweary Soap Operas: A Concealed Television Genre

genres

television

stories

Back in the 1980s I used to play with He-Man dolls. My favourite came pre-decomposed: Modulok and Multi-bot. Both were collections of interchangeable heads, torsos and limbs…

Part Ten: Log Likelihood estimation for Logistic Regression

statistics

R

Within this series, parts 1-4 formed what we might call ‘section one’, and part 5-9 ‘section two’.

Tidytuesday 2024-01-23

Educational attainment and town size

R

Education

Tidy Tuesday

This week’s TidyTuesday used data from the UK ONS which was explored in the 2023 article ’Why do children and young people in smaller towns do better academically than those…

Brendan Clarke, Andrew Saul, Nick Christofides, Kennedy Owusu-Afriyie, Jon Minton

The Other Left-Right Divide: Iain McGilchrist and the Battle of the Hemispheres

podcasts

stories

sci-fi

myth

In Watching The English, Fox (2005) writes that:

I got permanently banned from a politics forum for mentioning how circles work

politics

geometry

R

A month or so back, I was skimming a far-left-wing online politics forum, where a visual metaphor was presented to illustrate why it’s important that nominally left-wing…

Part Nine: Answering questions with honest uncertainty: Expected values and Predicted values

statistics

R

The last eight posts in this series have taken us into some fairly arcane territory, including concepts like the use of link functions and statistical families within GLM…

Part Eight: Guessing what a landscape looks like by feeling the curves beneath our feet

statistics

R

The previous post, perhaps the toughest of the series, showed how some special settings within R’s numerical optimisation optim() function can be used to estimate how much…

David Sedaris: Humourists as Unrepentant Observational Confessionals

books

mindfulness

comedy

humour

Whereas I devoted much of the 2022 Christmas break to (re?)learning Bayesian modelling, and over 2023 took four months off to learn software development, over the 2023…

Part Seven: Feeling Uncertain

statistics

R

In the previous post we managed to use numerical optimisation, with the optim() function, to good \(\beta\) estimates for linear regression model fit to some toy data. In…

Part Six: The Robo-Chauffeur

statistics

R

In the previous post in this series, I presented a function for calculating the log likelihood of a standard linear regression with a Normal error term. I then built a very…

Part Five: Traversing the Likelihood Landscape

statistics

R

The last part of the series ended by showing how, except for linear regression, the kinds of answers users need from models aren’t in tables of model coefficients, but in…

GLMs: My first series

I now have four fairly technical posts that form part of a series on understanding statistical modelling from a generalised linear regression (GLM) perspective. This series…

Part Four: why only betas just look at betas

statistics

R

This is part of a series of posts which introduce and discuss the implications of a general framework for thinking about statistical modelling. This framework is most…

Tidy Tuesday: Christmas films

A shorter and even tardier Tidy Tuesday this week, given we gave ourselves only half an hour rather than the usual hour to look over the most recent dataset.

Tom Fowler, Nick Christofides, Andrew Saul, Jon Minton

The Eerie Familiarity of Frasier (2023)

…And why it’s inverted Dr Who

stories

sitcoms

archetypes

mathematics

As with new Beavis and Butt-Head, which I’ve written about previously, Paramount+ includes access to new Frasier, a return of the vainglorious pratfalling public…

Changing tenure in Scotland

housing

As it’s something I’ve done already for work, but it’s all using public domain data, here’s a link to some charts showing how housing tenure in Scotland has changed over…

Nerdy Dialogues on Life and Death

Part 1: Introduction; Life Expectancy

demography

population health

methodology

Here’s an attempt to think some more about how and why standardised rates are used to compare populations. I’m doing so via the medium of a Socratic Dialogue¹.

Tidy Tuesday on Life Expectancy - Part Two

R

tidy tuesday

Life Expectancy

In the previous week’s TidyTuesday session, we looked at Life Expectancy.

Jon Minton, Andrew Saul, Nick Christofides, James McMahon, Kennedy Owusu-Afriyie, Sandra Nwobi

My Economic Inactivity Modelling Package: Informative Readme File!

R

Economic Inactivity

News

Packages

A few weeks ago, as I wasn’t using any personal or sensitive data, I decided to make the main repository where I keep my economic inactivity modelling work public, meaning…

Optimised for Twitter?

twitter

X

news

I’ve finally got around to ‘optimising the Quarto blog for Twitter’, by following the guidance in this section of Quarto’s website and adding the following declaration to the …

Why such pushback against 20 minute neighbourhoods?

research

walkability

driving

conspiracy theories

human development

I had the privilege yesterday of hearing a series of talks by researchers at the University of Glasgow on 20 minute neighbourhoods.¹ The talks covered areas like evidence…

Tidy Tuesday on Life Expectancy

R

tidy tuesday

Life Expectancy

This week’s Tidy Tuesday compares life expectancy across the globe and is available here:

Nick Christofides, Jon Minton, Sandra Nwobi

Beavis and Butt-Head: When a physics graduate dons Dumbface

cartoons

stories

eugenics

A second season of “Mike Judge’s Beavis & Butt-Head” is now available on Paramount+, continuing a series that began on MTV in the early 1990s. I’ve been watching it…

The Effective Savings on Interest-free Credit

statistics

finances

R

I have a Monzo account, and as part of the overall Monzo package make use of Monzo Flex, an interest-free credit service which means the payment made in one month is spread…

How to wrap presents

Christmas

Birthdays

Arts & Crafts

Here’s the key information I’ve learned from watching too many Youtube videos on how to wrap presents over the years.

Part Three: glm is just fancy lm

statistics

R

This is part of a series of posts which introduce and discuss the implications of a general framework for thinking about statistical modelling. This framework is most…

New blog feature: comments

quarto

blog

I think I’ve managed to set up a blog comment feature for each of the posts.

Part Two: Systematic components and link functions

statistics

This is part of a series of posts which introduce and discuss the implications of a general framework for thinking about statistical modelling. This framework is most…

Scientific Illustrations: Annotating the unit circle

statistics

R

graphics

economic inactivity

Here’s an example of a scientific illustration I’ve just produced to illustrate some scenarios I’m modelling for my work on economic inactivity determinants.

Interactive Sliders with Crosstalk and Plotly

R

plotly

crosstalk

Below is an example of creating a plotly chart with an interactive slider using crosstalk.

Tidy Tuesday on Dr Who

R

tidy tuesday

Dr Who

First we load the packages

Jon Minton, Nick Christofides

Robocop (1987) is wonderfully childish

stories

films

Here’s something that’s differently nerdy. Some thoughts on the enduring and childish appeal of Robocop as a character and concept, lifted largely from some notes I made on…

Part One: Model fitting as parameter calibration

statistics

This is part of a series of posts which introduce and discuss the implications of a general framework for thinking about statistical modelling. This framework is most…

Post with code

R

This short post is intended to confirm that I can run and render R code within a Quarto blog post.

EdinbR talk on modelling economic (in)activity transitions

R

modelling

talks

economics

health

Yesterday I had the great privilege of being one of two speakers at the Edinburgh R Users group, called EdinbR. (Difficult to say without sounding like a pirate.)

A Deathly Silence

Mortality

Epidemiology

Papers

What does it mean when someone dies, and no one notices for days, weeks, or months on end?

First Post

news

Hi, this is my first blog post. I’m making this using Quarto, starting off by slavishly following the tutorial, then incrementally adapting it to suit my preferences.