Blog

Longer, slower writing about research workflows, epidemiological methods, tools, and the craft of doing science.

This is where I write at length. Where Epi News is a regular surveillance digest, the blog is slower and more personal: how I organise projects, which methods I trust and why, the tools that survive contact with real work, and the occasional argument about how research should be done.

Posts are written in Quarto, so code, equations and figures are part of the text rather than screenshots of it. You can subscribe to the RSS feed, filter by category, or search the archive below.

A Markdown-First Research Workflow for My Colleagues (and Their AI Agents)

AI
workflow
Markdown
Quarto
Git
reproducibility

A practical guide to the research setup I use: Markdown and Quarto, project-level bibliographies, Git, the university OneDrive, WSL, getbib, and a disciplined way for people and AI agents to work in the same project.

27 August 2026
20 min

AI as the SPSS, R, and Stata for Qualitative Research

AI
qualitative research
workflow
methodology
reproducibility

Quantitative researchers spent forty years learning to delegate computation to software without surrendering judgment. Qualitative researchers now face the same transition, with a tool that answers in the register of analysis rather than in numbers. A delegation framework, a project structure, and a reading of where the methodological dispute currently stands.

20 August 2026
29 min

A Package That Argues With You: kkstatfun, One Analysis Start to Finish

R
kkstatfun
epidemiology
causal inference
public health
reproducibility
workflow

One cohort, one deliberately misleading crude estimate, and a worked tour of kkstatfun — from 2×2 tables and colour-blindness audits to targeted causal inference, interrupted time series and cost-effectiveness curves. Then the public-health workflow it sits inside: getting NHIS data out of the system under ЗДОИ, the pipeline that cleans it, where AI actually helps, and how it reaches a manuscript.

16 August 2026
38 min

The PhD Curriculum Is Preparing Researchers for Yesterday

teaching
doctoral education
reproducibility
workflow
opinion

Our doctoral school allocates 40 hours to SPSS and three to epidemiology, and mentions AI exactly zero times. A constructive argument — with worked examples in Markdown, Quarto, git and the terminal — for what a modern doctoral curriculum should add, and a concrete mechanism for adding it.

22 July 2026
28 min

My Academic Writing Setup: Why Writing Has Already Changed

workflow
reproducibility
R
open science
writing

Academic writing is not about to change — it changed while we were arguing about whether it would. Here is how I organise projects, references, analyses and tools now that producing clean prose is no longer the bottleneck.

20 July 2026
28 min
No matching items