UC Carpentries: Python (JupyterLite fallback)

A browser-based Python environment for UC Carpentries Python workshops. It runs entirely in the browser using JupyterLite: no install, no account, no setup. It exists as a fallback for learners who can't install Python locally (locked-down laptops, install trouble during a workshop, or wanting to start fast).

It covers both Python lessons UC Carpentries teaches:

• Library Carpentry: Introduction to Python (lc-python-intro)
• Software Carpentry: Plotting and Programming in Python (python-novice-gapminder)

Live site: https://ucla-data-science-center.github.io/uc-carpentries-python-lite/

Both lessons' data and a welcome notebook are bundled in, so file paths like data/2011_circ.csv and data/gapminder_gdp_oceania.csv work exactly as the lesson text writes them.

---

For learners

1. Open the live site: https://ucla-data-science-center.github.io/uc-carpentries-python-lite/. It loads JupyterLab in your browser. First load can take a few seconds while the Python kernel downloads.
2. The welcome.ipynb notebook opens automatically. Run its first code cell to confirm everything works.
3. To start a fresh notebook: click + at the top of the file browser, then choose Python (XPython).
4. Follow your lesson:
   • Library Carpentry: https://librarycarpentry.github.io/lc-python-intro/
   • Software Carpentry: https://swcarpentry.github.io/python-novice-gapminder/

Both lessons' datasets live together in data/: the 12 Chicago Public Library circulation CSVs (2011-2022) plus all_years.pkl / df_long.pkl for Library Carpentry, and the Gapminder GDP CSVs for Software Carpentry.

What's different from a normal Jupyter install

• No shell commands. Lines starting with ! (such as !pip install or !ls) don't work in JupyterLite. Everything you need is preinstalled.
• Preinstalled packages: pandas, numpy, matplotlib, seaborn, plotly. If a lesson asks you to install something else, it won't be available here.
• Your work lives in the browser. Notebooks you create are saved to your browser's local storage, not to a server. Clearing site data wipes them. Download anything you want to keep (right-click the file → Download).

---

For maintainers

This repo is a JupyterLite site that GitHub Actions builds and publishes to GitHub Pages on every push to main.

Layout

Path	Purpose
content/welcome.ipynb	The notebook that opens on load.
content/lc-python/data/	Library Carpentry datasets.
content/gapminder/data/	Software Carpentry datasets (copied from swcarpentry/python-novice-gapminder).
environment.yml	The in-browser kernel and its packages (xeus-python + the scientific stack), resolved from emscripten-forge.
requirements.txt	The build-time tools that run in CI (jupyterlite, jupyterlite-xeus, jupyter-server), version-pinned.
jupyter-lite.json	JupyterLite config. Sets the notebook that opens on load (welcome.ipynb).
.github/workflows/deploy.yml	Build + deploy to Pages.

The build passes one --contents flag per lesson. Each lesson keeps its data in its own data/ subfolder, and those subfolders merge into a single flat data/ in the built site, so lesson-hardcoded paths work without a prefix.

Build locally

pip install -r requirements.txt
jupyter lite build \
  --contents content/welcome.ipynb \
  --contents content/lc-python \
  --contents content/gapminder
# serve the result
python -m http.server -d _output 8000   # then open http://localhost:8000

_output/ is git-ignored; it's a build artifact.

See MAINTAINING.md for how to update the data, add a lesson, change packages, edit the welcome notebook, and troubleshoot the deploy.