Course Overview
Course Overview
Health Data Science: AI and Knowledge Translation is a beginner-friendly, open course that teaches you to work with real health data using R, Quarto, Git, GitHub Codespaces, and AI coding assistants — and to communicate results clearly, reproducibly, and responsibly.
This is an open educational resource (OER). You can follow it independently, or an instructor may adopt it for a specific course. If you are taking this in a specific offering, that course’s Canvas (or equivalent) is the source of truth for the schedule, deadlines, grading, and your instructor’s contact details. This overview intentionally leaves those out so the material stays useful to any learner.
What you’ll learn
By working through this course you will be able to:
- Explain key ethical, privacy, security, Indigenous Data Sovereignty, and knowledge-translation issues in health data science, including responsible use of generative AI tools.
- Use GitHub Codespaces, Quarto, Git, and GitHub to organize, document, version, and reproduce a small health-data analysis.
- Import, clean, summarize, and visualize health-related data using R and the tidyverse.
- Read and compare short Python/pandas examples with AI assistance, recognize common R↔︎Python translation traps, and explain when Python may be the more appropriate tool.
- Use generative AI as a coding assistant while critically auditing its output for coding errors, hallucinated claims, bias, security risks, and reproducibility problems.
- Conduct a bounded exploratory data analysis and communicate descriptive findings using clear tables, figures, narrative interpretation, and appropriate limitations.
- Create a dashboard-style knowledge-translation product and a reproducible Quarto report for a defined audience.
- Use a shared devcontainer and dependency documentation so analyses can be rerun in a fresh Codespace.
- Give constructive peer feedback on the clarity, reproducibility, auditability, and knowledge-translation value of a project.
Who it’s for
No prior experience required. This is a “Zero to AI Co-Pilot” course: you do not need previous experience with R, Python, statistics, or programming. Comfort with a web browser, basic file management, and a willingness to troubleshoot are assumed; everything else is taught from the ground up.
How the course works
- Cloud-first. All coding runs in GitHub Codespaces — a cloud computer you reach from any modern web browser. No local installation is required.
- R is the core language, with Quarto for reproducible documents; short Python/pandas examples build polyglot awareness.
- AI as a co-pilot, always audited. You use LLMs to draft, debug, and translate code, then check their output against your data and the documentation.
- Reproducible by default. Work is version-controlled with Git and designed to re-run in a fresh environment.
What the course is built around
- The Full Data Lifecycle — from ethical data access and cleaning to secure cloud computing and publication.
- Reproducible Science — Git-based workflows, a shared environment, explicit dependency notes, and literate programming in Quarto.
- AI-Augmented Workflows — using LLMs as coding co-pilots while auditing their output for errors, hallucinations, bias, security risks, and reproducibility failures.
- Knowledge Translation (KT) — the “last mile”: turning results into dashboard-style products, visual stories, plain-language summaries, and policy-ready reports.
- Real-World Health Data — hands-on work with messy, real datasets (for example, NHANES) rather than toy examples.
Weekly topics
The material is a weekly progression; each week pairs a short concept seminar with hands-on studio work.
| Week | Topic |
|---|---|
| 0 | Onboarding: accounts, orientation, and setup (self-paced) |
| 1 | Health data, knowledge translation, and ethics |
| 2 | Modern research workflows: Codespaces, Quarto, Git |
| 3 | R with AI: data, functions, the tidyverse, debugging |
| 4 | Git and collaboration: commits, branches, pull requests |
| 5 | Polyglot awareness and R deepening (Python/pandas) |
| 6 | Data visualization and AI-assisted visual audits |
| 7 | Exploratory analysis, Table 1, and AI auditing |
| 8 | Dashboard prototypes for knowledge translation |
| 9 | Communicating scientific findings |
| 10 | Writing and publishing reproducible reports |
| 11 | Portfolio surgery: reproducibility stress-testing |
Use the sidebar to work through each week.
About the developer
This open course was developed by Dr. M. Ehsan Karim (School of Population and Public Health, University of British Columbia), with the AI-KT team (Manya Jain, Rainie Fu, Md. Belal Hossain). It is shared as an open educational resource for learners and adopting instructors.