Econometrics Workbench

An interactive self-study site for econometrics: intuition first, live R apps, then the maths with every symbol translated. Everything runs in the browser (Shinylive + webR), so the site is plain static files: no server, no R needed by students.

What's inside

index.qmd, about.qmd          home page and "how to use this site"
lessons/                      3 lessons, each with Shinylive apps (8 apps in total)
  01-regression.qmd             least squares, confidence intervals, omitted variable bias
  02-likelihood.qmd             coin likelihood, Challenger O-rings, sampling distribution of the MLE
  03-bootstrap.qmd              bootstrap machine, when the bootstrap fails
labs/                         3 labs with live, auto-checked R exercises (16 in total)
styles/theme.scss             the visual design
_extensions/                  quarto-ext/shinylive and r-wasm/quarto-live (vendored, don't edit)
.github/workflows/publish.yml builds and publishes to GitHub Pages on every push

Quickest way to put it online (no installation)

The econ-workbench-site.zip download is the finished website. Unzip it and drag the folder onto https://app.netlify.com/drop, or upload its contents to any static web host. It must be served over http(s); opening index.html straight from your disk won't start the R engine.

To preview it locally, run this inside the unzipped folder and open http://localhost:8000:

python3 -m http.server 8000

Editing and rebuilding

You need Quarto (1.5 or newer) and R with these packages:

install.packages(c("shinylive", "shiny", "bslib", "knitr", "rmarkdown"))

Then from this folder:

quarto preview      # live preview while you edit
quarto render       # build the site into _site/

Publishing with GitHub Pages

1. Put this folder in a GitHub repository (the .gitignore keeps _site/ out).
2. Run quarto publish gh-pages once from your computer. This creates the gh-pages branch.
3. In the repository settings, set Pages to deploy from the gh-pages branch.
4. From then on, every push to main rebuilds and republishes automatically via the included workflow.

Adding a new topic

Copy a lesson and its lab, and add both to the sidebar in _quarto.yml.

• Apps are {shinylive-r} blocks with #| standalone: true. Each is a complete Shiny app. Develop it in RStudio first with shiny::runApp(), then paste it in. Keep to base graphics, shiny and bslib where you can: every extra package adds download time for students.
• Exercises are {webr} blocks. An exercise has up to four parts sharing one exercise: id: a setup: true block, the student's block (with ______ blanks), a check: true block, and .hint / .solution divs. The check block sees the student's last value as .result and their environment as .envir_result, and returns list(correct = TRUE/FALSE, message = "...").
• Styling helpers used in the lessons:
  • ::: {.bench .column-page-right} frames an app on graph paper at full width.
  • ::: {.try-this} is the challenge list under each app.
  • ::: {.decoder} is the symbol-by-symbol formula table; start it with [Formula decoder]{.decoder-title}.
  • [text]{.scribble} in a ::: {.column-margin} block is a handwritten margin note (add .blue for blue ink).
  • <span class="key">…</span> is the highlighter for the one sentence that matters most.

Notes

• The first app or code cell on a page takes a few seconds to load R (around 20–30 MB, cached afterwards).
• Content follows the lecture notes Regression basics, Maximum likelihood and The bootstrap by Lukáš Lafférs (Matej Bel University). The footer credits them; if you publish publicly and aren't the author, check with him first.