Fresh Codespace Test
Goal
Test whether a project can run in a fresh environment and document the results and any failures before final submission.
This page is the main Week 11 runbook. Use it for the Portfolio Surgery session, M4 peer review, and final portfolio preparation.
Roles
For group testing, assign roles:
- Driver: runs the project exactly as documented.
- Navigator: reads instructions aloud and watches for ambiguity.
- Auditor: logs paths, dependencies, citations, data provenance, and AI-use issues.
Rotate after the first major failure or after 20 minutes.
Runbook
Complete these steps in a new Codespace or equivalent clean environment. Work from the repository root unless the project README says otherwise.
- Open the repository from GitHub.
- Record the latest commit hash.
- Read the README before running code.
- Follow the documented setup steps only.
- Restore or install dependencies using documented instructions.
- Render the final report or book.
- Open or run the dashboard-style KT product.
- Check that data paths are relative.
- Check that citations resolve.
- Check that figures and tables are generated from code.
- Check that AI-use notes are present and specific.
- Record the check results and any failure, warning, unclear instruction, or repair.
- Commit any repairs with clear messages.
- Submit the updated repository link and commit hash on Canvas.
Failure Log
Use this table for the Portfolio Surgery session and for peer review.
| Check | Result | Evidence | Fix made | Owner |
|---|---|---|---|---|
| Repository opens in fresh environment | ||||
| Packages install or load | ||||
| Report renders | ||||
| Dashboard or KT product opens | ||||
| Data paths are relative | ||||
| Citations resolve | ||||
| README instructions are accurate | ||||
| AI-use note is complete |
If every check passes, keep the evidence for each check and state that no failures were identified and no repairs were required. You will not lose marks for having no failures to repair. Do not create an artificial failure or make an unnecessary change.
Example Failure Log
| Check | Result | Evidence | Fix made | Owner |
|---|---|---|---|---|
| Repository opens in fresh environment | Pass | Codespace opened from GitHub | None | Group |
| Packages install or load | Needs repair | library(tidyverse) failed |
Added dependency note to README | Analyst |
| Report renders | Pass | report.html created |
None | Writer |
| Dashboard opens | Needs repair | data path failed from repo root | Replaced absolute path with relative path | Developer |
| Citations resolve | Pass | reference list appears | None | Writer |
NHANES Exemplar Checks
For the NHANES case study, the checks are:
- load
examples/nhanes-equity/data/nhanes_equity_v6.csv; - render the Week 10 report template;
- confirm the README explains cached data and optional CDC retrieval;
- if using the Shiny exemplar as a demo, launch it from the repository root;
- if using the static dashboard demo, open
weeks/week11-portfolio-surgery/static-dashboard-ojs.qmd.
The CDC rebuild and static dashboard are optional/demo pathways, not required portfolio work.
Optional Python Pathway
If the project uses Python, add a short dependency note with:
- how Python is launched;
- required packages;
- whether a notebook, script, or Quarto Python chunk is the primary workflow;
- how outputs are regenerated.
Keep the Python pathway documented, but do not add it unless it supports the core audience-facing product.