EdinbR: The Edinburgh R Usergroup

Thu, Nov 20 · 5:00 PM GMT

Nov 2025: Causal modelling and Word docs from Quarto

Appleton tower, Crichtin Street, Edinburgh, al, GB

* Title: Nov 2025: Causal modelling and Word docs from Quarto
* Date: Thursday 20 November 2025, 5.00PM - 6.00PM
* Location: 2.14 Appleton Tower, EH8 9LE
---

Stephanie Droop: Extensions and issues in causal modelling

Stephanie Droop has almost finished a PhD in computational cognitive modelling at the University of Edinburgh. She has a background in cognitive psychology and studies how people think about causes and explanations.

The human mind has a strong intuitive sense of when something is a cause of something else. These patterns have traditionally been studied in simple urn scenarios and in artificially separated tasks. I extend the notion to richer and more complex scenarios and on the way give a whistlestop tour of issues in causal modelling.

---

Jon Minton: Building styled scentific reports (for them and us) using modularised Quarto documents

Jon Minton is a trial statistician with a background in public health, interested in population structure, data science, and more recently software development. He has two cats who occasionally write code and emails for him when they walk over his keyboard.

When working in any organisation, there tend to be standards for how documents should look, including company colours, fonts, logos and so on. By contrast, most data scientists and statisticians are *just* interested in the accuracy of the data, appropriateness and detailing of the methods, and the results. All too often, the transferring and styling of materials that matter to 'us', into corporate formats that look professional to 'them', is a manual and error-prone process, done by in-house copy-editors and then reviewed by middle managers and domain experts. And if queries, errors and requests relating to the data, methods and results are raised after 'our' materials are styled to look professional, the whole manual process needs to start again.
Quarto is Posit's (formerly RStudio's) attempt to evolve and generalise RMarkdown (itself borrowing conceptually from Jupyter notebooks, and of course markdown) to work with other languages, and crucially to produce a much wider range of possible outputs, beyond HTML web pages.
Used wisely, Quarto offers the means to bridge the gap between the expectations of 'us' (data scientists and statisticians) and 'them' (other people in an organisation) in a way that's much more automated, and so quicker, less error-prone, and easier to iterate over.
Using Quarto this ways relies on two important features of this type of document: firstly the ability (through pandoc) to generate styled word documents based on an appropriately formatted word document to use as a style template; secondly the ability for Quarto documents to inherit (through the include keyword) and render other Quarto documents, allowing scientific materials to be worked on and re-organised in a highly flexible and modular framework.
This presentation will introduce the key principles and ideas for how to organise and work with Quarto files this way, showing how modularity, styling and structuring can be implemented in practice. As well as showing how the same materials can be reorganised and rendered into different styled word documents for different non-technical audiences, it will also show how the same approach can be used to produce interactive web pages that facilitate more technical reviews by internal peer reviewers, without requiring them to rerun our analyses directly.