Repository navigation
Open
0 / 40 of 4 issues completedLabels
Description
This is a big-picture plan for cie, written after reading the source at bb27598, the 30 open issues and the 3 open pull requests.
It is a plan only: nothing here has been started.
Each workstream below is a sub-issue, and the existing issues sit under the workstream they belong to.
Where the repo stands
- The book has about 570 lines of prose and code outside the macros file.
Five chapters render: the preface, an introduction to causal inference, collider bias, bias analyses and difference in differences. - Two rendered pages are still the Quarto template's placeholders.
summary.qmdtells readers "this book has no content whatsoever", andintro.qmd(not rendered) is the template's literate-programming stub. bias-analyses.qmdis one long quotation, andcollider-bias.qmdrests on a single web source.dnd.qmdandadnd.qmdhold difference-in-differences material that no chapter list includes.- 19 of the 30 open issues ask for a link or a blurb for an outside resource.
The preface carries the existing resource lists as hand-written Markdown. - The configuration was copied from
rmeand still says so:repo-urlpoints atd-morrison/rme, the HTMLsubjectis "Epidemiology 204", andDESCRIPTIONhas theusethisplaceholder title and description. macros.qmdis a 389-line local copy of the lab macros rather than theMorrison-Lab/macrossubmodule.publish.ymlandpreview.ymlare hand-written and pin actions at@HEADand@main; the checks added since October already callMorrison-Lab/gha@v3.
Workstreams, in the order I recommend
- Foundations.
Bring the configuration, macros, CI and agent instructions in line with the lab's other notes sites, and convert the book to a website.
This is small and unblocks everything else. - Resource catalog.
Move the resource lists into one structured file that the pages read, then add the 19 requested resources to it.
A handful of small pull requests closes most of the backlog. - Core chapters.
Settle the outline, replace the placeholder and stub chapters, and write the foundation chapters the later ones depend on. - Difference in differences.
Merge the three DiD files into one chapter that renders.
Decisions for Ezra
- How do
cieandwindivide the subject?
I recommend thatwinstays the chapter-by-chapter companion to What If, andcieholds the lab's own topic notes and the methods the book does not cover (difference in differences, quantitative bias analysis, competing events).
ciethen links towin's definitions instead of restating them, because both sites already publish a dependency graph. - Book or website?
I recommend the website, as #32 already asks, because it gives each chapter slides and a handout.
Open pull requests
- #53 adds an optional philosophy of causality chapter. It belongs to the core chapters workstream.
- #40 points agents at ai-config's rules. It belongs to foundations.
- #30 is a Copilot draft from April that writes the competing-events summary into
summary.qmd. I recommend closing it and writing that summary as its own page under core chapters.
Posted by Claude Code (AI agent) --- not written by a human.
Activity