- Inputs, code, dictionary and environment record.
- Consistent numbers, tables and figures.
- README with execution and export checks.
The computational chain
Inputs lead to checks and analysis, then numbers, tables, figures and prose. Inline computation lets a changed input update manuscript values as well as tables. This reduces manual inconsistency, not methodological error.
A thesis still needs a question, argument, justified methods and human interpretation. Begin with institutional formatting; generating a PDF does not establish acceptance of its layout.
The teaching project
The download contains entirely synthetic data, standard-Python analysis, README, expected outputs and a QMD manuscript. Paths are relative and there are no credentials. Separate inputs from regenerated outputs and retain the dictionary.
Run python research.py first. It creates a teaching summary, six-path table and SVG without external statistical packages. Quarto rendering additionally needs Quarto and the Jupyter environment described in the README.
Connect numbers to prose
A Python cell calculates mean_score; the inline expression {python} mean_score inserts it into a sentence. Avoid typing the same number manually. Reconcile figure and table identifiers with the exported document.
Change a score in a teaching copy, rerun from the beginning and confirm the mean, table, plot and sentence use the same release. Preserve a dataset hash or release identifier.
Clean execution and environments
Execute in a fresh environment and documented order. Record Python, Quarto and package versions; a random seed alone is insufficient. Cached outputs can become stale, so specify the actual rebuilding command.
Check counts, ranges, units, transformations, analytical sample and known-answer cases independently. The example is descriptive, not a universally validated thesis method.
RTL, Word and PDF
For HTML, separate page direction from LTR code and DOI text. Word requires appropriate reference-document styles and paragraph direction. PDF requires a suitable engine and language-capable fonts, tested in the real export.
Inspect contents, numbering, footnotes, bilingual citations, wide tables and page breaks separately in each format. If Word is edited manually afterward, document those edits and the regeneration procedure.
Completed teaching worksheet
This is a hypothetical teaching case, not observed data, an actual review or a publication acceptance. Numbers illustrate decisions.
| Decision or record | Teaching example | Your project action |
|---|---|---|
| Data | 12 synthetic teaching records | Define real rights, release and dictionary. |
| Analysis | Standard-Python research.py | Start with a clean run. |
| Prose | mean_score from a computation cell | Connect text values to outputs. |
| Check | Change an input and rerun | Reconcile sentence, table and figure. |
| Export | Example HTML and editable QMD | Review institutional Word/PDF separately. |
Deliverables and completion checks
- Inputs, code, dictionary and environment record.
- Consistent numbers, tables and figures.
- README with execution and export checks.
Version and scope references: Quarto — Inline computational results · Quarto — Word output
Sources and further reading
Official sources for verification and further reading
- Quarto — Inline computational results Source verification: 2026-10-06
- Quarto — Word output Source verification: 2026-10-06
- Quarto — PDF output and engines Source verification: 2026-10-06
- Cornell Data Services — Writing READMEs for Research Data