Programming in Python for Data Science

This repository holds the Quarto source for Programming in Python for Data Science, a UBC Extended Learning's Key Capabilities in Data Science program course.

The rendered course is published at https://ubc-mds.github.io/programming-in-python-for-data-science/.

The course teaches data analysis in Python, including:

• Processing tabular data with pandas.
• Visualizing processed data with altair.
• Working with numerical data using numpy.
• Iteration, flow control, function design, testing, and debugging skills.

It assumes no prior programming experience.

Every exercise runs in the browser through Quarto Live and Pyodide, so learners need nothing installed to take the course — the setup below is only for people editing the material.

Course Instructors

• Socorro Dominguez Vidana

Former Course Instructors

• Celine Siliang Liu

Contributors

This course was built and is maintained by many people from UBC MDS program.

• Tomas Beuzen
• Hayley Boyce
• Mike Gelbart
• Tim Head
• Varada Kolhatkar
• Joel Ostblom
• Tiffany Timbers
• Elijah Willie
• Mengxin (Betty) Zhao

Attribution

• Material adapted from UBC's DSCI 511: Python Programming for Data Science by Tom Beuzen.
• The cereal dataset: 80 Cereals (c) by Chris Crawford, licensed under CC BY-SA 3.0.

Tips for Contributing

Issues and bug reports are always welcome — please open an issue.

For changes, branch off master and open a pull request. Every pull request automatically gets a preview deployment of the full site (see Deployment), so you can check how your change renders before it is merged. Because the material is course content, prose edits matter as much as code edits:

• Keep the conversational voice of the surrounding text
• Slide speaker notes come from the videos' scripts.

A previous version of the course material was developed using Ines Montani Spacy's teaching framework.

What this repository contains

The course is nine modules plus a bonus module and a closing section:

Module	Topic
M0	Welcome to Programming in Python for Data Science
M1	Python & Pandas — An Unexpected Friendship
M2	Not So Scary Wrangling (Table Manipulation and Chaining)
M3	Tidy Data and Joining Dataframes
M4	Python Without the "Eek" (Basic Python)
M5	Making Choices and Repeating Iterations
M6	Function Fundamentals and Best Practices
M7	Importing Files and the Coding Style Guide
M8	A Slice of NumPy and Advanced Data Wrangling
Bonus	Bonus module
—	Module closing remarks

Each module directory holds the module's pages, and a slides/ subdirectory holding its reveal.js decks.

Repository layout

Path	What it is
_quarto.yml	Project config: render list, resources, and the hand-maintained sidebar
modules/	All course content — the only directory that is rendered
data/	CSV/Excel datasets used by lessons and exercises
src/utils.py	Shared Python helpers: display settings, Altair theme, exercise checkers
src/quiz.js	The generateQuiz() multiple-choice widget
static/	Images, logos, and other media
styles.scss	Site and slide theming
_extensions/	Vendored Quarto Live extension
docs/	Syllabus and the video screen-recording workflow

In previous git commits (Up to commit c363d86fb411dbfc7e2dbcbd38cee62d7aac8984), you can find the folders chapters/, exercises/, and junk/. These are legacy content from the course's pre-Quarto platform. They are not built and not maintained but can be used as a reference for the original exercise numbering.

How it works

_quarto.yml drives everything:

• execute-dir: project — code cells resolve paths from the repository root, so a lesson reads data/canucks.csv, never a path relative to its own .qmd.
• render: includes only modules/**/*.qmd. Anything outside modules/ is not built.
• resources: copies styles.scss, src/utils.py, and data/ into the site unexecuted. A separate top-level pyodide: block copies the same files into the browser's virtual filesystem so in-page code can read them.
• The sidebar contents: list is written by hand. A new page will not appear on the site until it is added there.

Pages come in three shapes:

1. Video and slides — format: html, a tabset holding a YouTube embed and an <iframe> pointing at the page's rendered reveal.js deck.
2. Quiz — format: html, loading src/quiz.js and calling generateQuiz() once per question.
3. Live coding exercise — format: live-html, using {pyodide} cells that execute in the learner's browser. A cell tagged #| exercise: <slug> holds the fill-in-the-blank starter; a second cell with the same slug and #| check: true asserts against the learner's result and reports back through the helpers in src/utils.py. Matching .hint and .solution blocks are keyed by the same slug.

Running the site locally

System requirements

• Quarto 1.6 or newer (1.6.43 is what CI uses)
• Python 3.11 or 3.12
• The packages listed in environment.yaml (pandas, altair, jupyter, openpyxl, and others)

Setup

git clone git@github.com:UBC-MDS/programming-in-python-for-data-science.git
cd programming-in-python-for-data-science
conda env create -f environment.yaml
conda activate kcds-prog

The environment is named kcds-prog and pins Quarto itself, so activating it gives you a matching Quarto.

Build

quarto preview   # local server with live reload — the usual way to work
quarto render    # full static build into _site/

To rebuild a single page instead of the whole site:

quarto render modules/module1/module1-01-introduction_to_dataframes.qmd

A full quarto render executes every reveal.js deck through Jupyter and takes a while, so prefer quarto preview or a single-file render while editing.

Troubleshooting

Rendering any page starts a local Jupyter kernel — including live-html pages, whose {pyodide} cells run in the browser rather than locally. No .qmd pins a kernel with a jupyter: key, so Quarto picks the first Python kernelspec it finds. If a stale kernelspec on your machine points at an interpreter that no longer exists, the render fails with a FileNotFoundError naming a path that has nothing to do with this repository. Check with:

jupyter kernelspec list

and remove or repair any entry pointing at a deleted environment.

_site/ and .quarto/ are build output. Neither belongs in a commit.

Deployment

Workflow	Trigger	Result
.github/workflows/publish.yaml	Push to master	Renders the site and pushes _site/ to the gh-pages branch
.github/workflows/pr-preview.yaml	Pull request against master	Builds and deploys a preview, removed when the PR closes

Both build the environment from environment.yaml with micromamba and run quarto render, so a change that renders cleanly for you locally will render in CI.

License

This repository carries two licenses, split between the instructional material and the software that frames it:

• Instructional material — the course content — is licensed CC BY-NC-SA 4.0.
• Software — the framework used to build and host the material — is licensed under the MIT License.

See LICENSE.md for the full text of both.