The R Project

Week 10 — composing your Module B skills into one short focused R-in-Quarto report on a question of your choice

A short conceptual reading on the R Project, the project that closes Module B — R / computation, visualization, simulation, reporting. The Week 10 schedule slot is studio time for the project, and the Week 10 R Project conference is the required term-anchor conference for the week. The exact project prompt, rubric, deadlines, submission details, extension procedures, and conference sign-up mechanics live in the Assignments/LMS space.

Week 7 opened Module B with the first R-in-Quarto report on a small tidy dataset. Week 8 added one focused ggplot2 figure inside the same Quarto container. Week 9 added a reproducible simulation to the same container, with set.seed() as the load-bearing function. Week 10 is the week where you put those three substances together into one short focused report on a question of your choice.

The R Project is not a new-skills week. There is no new editor, no new render engine, no new portfolio convention, and no new package install required. Everything the project asks for is already in the kit you built across Weeks 7–9: R chunks inside a Quarto-to-PDF report (Week 7), one focused ggplot2 figure with labels and prose interpretation (Week 8), and a small reproducible simulation under a stated seed (Week 9). What changes in Week 10 is composition — you choose the question, choose the track, scope the work to a short report, defend the choices at the midweek conference, and ship.

The R Project sits in your portfolio next to the LaTeX Project from Week 6 — both are short, polished, reproducible artifacts that compose a module’s skills into one document.

One container, three substances, one project

Modules A and B share the same Quarto-to-PDF render chain. Week 1 set up the render chain; Weeks 2–4 added math notation, structure, and apparatus; Weeks 5–6 produced the LaTeX Project. Week 7 reopened the container for R; Week 8 added a figure; Week 9 added a simulation. Week 10 asks one question — what does it look like to write a short focused report that uses one or two of those substances on a question you care about?

The deliverable is the same shape as your weekly Week 7, 8, or 9 reports — a single Quarto-to-PDF rendered document a few pages long, structured as headings + chunks + prose, with an AI Use Note at the end. The difference is that you choose the question, you choose the data or the simulation, and you have a midweek conference to confirm the choice before submission.

The R Project is non-droppable

The weekly drop pool (best 9 of 11) covers HW 7, HW 8, HW 9, and most other weekly assignments. The R Project is not in that pool. It is the major artifact for the R / computation / visualization / simulation / reporting category of your grade, and it cannot be skipped — exactly the same role the LaTeX Project played for the LaTeX / mathematical-writing category.

Project extensions require a request submitted before the due date, granted at the instructor’s discretion, per the syllabus. The weekly flat-20% / one-week-late rule does not automatically apply to the project. The exact request procedure lives in the course LMS.

The two tracks

The R Project has two named tracks. You pick exactly one. The Week 10 R Project conference is where the track choice gets confirmed.

Track A — Data analysis and visualization with ggplot2

You pick a small dataset and a small question about it, inspect the data, compute summaries that bear on the question, build one focused ggplot2 figure that addresses the question, and write prose interpreting what the rendered evidence shows.

This is the R Project track that extends Week 8 most directly. If Lab 6 and your Week 8 assignment felt like the work you want to do more of, Track A is probably your shape.

Track B — Simulation, sampling behavior, or a CLT-style investigation

You pick a small random process and a small question about it, set a seed with set.seed(), simulate the process reproducibly, summarize the simulated outcomes, optionally visualize the simulated distribution, and write prose interpreting what the simulation under the stated seed shows — without overclaiming about probability theory, populations, or other seeds.

This is the R Project track that extends Week 9 most directly. If Lab 7 and your Week 9 assignment felt like the work you want to do more of, Track B is probably your shape.

You do not have to do both tracks. A project that tries to do Track A and Track B in one report is a scope problem — the report loses focus and neither track gets finished cleanly. Pick one.

Picking a track — an honest selection guide

There are three useful questions for picking a track:

  • Which Module B lab did you enjoy more — Lab 6 or Lab 7? Lab 6 → Track A. Lab 7 → Track B.
  • Which Module B weekly did you feel more confident on — the ggplot one or the simulation one? ggplot → Track A. Simulation → Track B.
  • Do you have a question about a real dataset you have been curious about, or about a small random process you have been curious about? Dataset → Track A. Random process → Track B.

If after these three questions you are still unsure, bring the question you most want to answer to the Week 10 R Project conference and we can pick the track together at the conference. That is part of what the conference is for.

Track A — what data is allowed?

For Track A, use a built-in R dataset, a small public dataset with verifiable provenance, or a student-selected dataset approved through the course process. Simulated data with set.seed() belongs to Track B, not Track A — a Track A project analyzes data that already exists; it does not generate its own data.

  • Built-in R datasets (mtcars, iris, airquality, Orange, ToothGrowth, women, cars, trees, and other built-ins) are always allowed. You have already used some of these in Weeks 7 and 8. They are not a downgrade — a tight Track A project on mtcars or iris is a complete project.
  • Small public datasets with verifiable provenance are allowed. Include a citation or URL the reader can resolve, and keep the data file inside your project folder so the document renders reproducibly.
  • Student-selected datasets approved through the course process are allowed; bring the candidate to conference and the exact approval mechanics live in the course LMS.

The general data policy for the course is on the Data guidelines page; Track A draws from the three options above.

You do not need to find an external dataset to do Track A. A focused analysis of a small built-in dataset is a complete Track A project.

Track B — what reproducibility looks like

Track B’s load-bearing skill is the same one from Week 9: call set.seed(N) once, before any random call, in a chunk that runs first. Without that, the report’s numbers change on every render and the prose interpretation drifts away from the rendered output silently.

State the integer seed value explicitly in your introduction paragraph as well as in the chunk — for example, “with set.seed(2026)” — so a reader can re-run with the same seed and get the same numbers.

Track B does not require a formal Central Limit Theorem statement, a proof, a hypothesis test, a confidence interval, a standard-error formula, or a Monte Carlo error analysis. It asks for a small reproducible simulation, an honest summary, and a careful interpretation under the stated seed. Sampling-behavior extensions with replicate() are welcome but optional.

A polished short report, a few pages

The target on both tracks is a polished short report of a few pages. The same calibration that worked for the LaTeX Project works here: one focused piece of rendered evidence (one summary + one figure for Track A; one simulation + one summary for Track B), a clear prose interpretation around it, and a clean PDF render. Not five plots. Not three datasets. Not a sweeping discussion section.

The pull toward the end of the week is to add more — a second dataset, a second ggplot, a second seed, a third distribution to simulate, a long Discussion. That turns a polished short report into an unpolished long one. Finish the one focused report you planned at conference.

The Week 10 R Project conference

The Week 10 R Project conference is the required term-anchor conference for the week. It is one of the five required conferences (Weeks 1, 4, 7, 10, 13) — see the Syllabus for the full list. It is part of the project, not a separate event.

  • When. Early in Week 10. Exact sign-up slots live in the course LMS.
  • What it confirms. Per the syllabus: your project track, dataset or simulation plan, and early report structure before the project submission.
  • What to bring.
    • A draft Quarto source with at least the title block, the track declaration, the intro paragraph, and rough section headings.
    • Your dataset (Track A) or the seed value plus simulated process you plan to run (Track B).
    • A one-sentence answer to “what is this project trying to show?”
  • What the conference is for. A short project-go/no-go. If the track or scope is not viable in the time left, the conference is the moment to course-correct. After the conference you have the rest of the week to finish.

If the time you signed up for stops working, reschedule inside the same week through the course LMS. The exact sign-up and rescheduling procedures live there.

Finishing well — debugging hints

Before you submit, render the project twice in a row and confirm the numbers and figures are the same in both renders. This catches the most common Week 10 problems before the LMS grader does.

  • The render fails. Look at the first error line; comment out the failing chunk; render; isolate the problem. Bring the error message to a scheduled studio meeting in the MAC for help.
  • The numbers change on the second render (Track B). set.seed() is missing, or is being called after the random call instead of before. Move it to the top of a setup chunk that runs first. This is the Week 9 failure mode, applied to the project.
  • The numbers or the figure change on the second render (Track A). A Track A analysis should produce the same rendered output on every render of the same dataset. Confirm the data file is inside your project folder so the path is stable, and confirm you are loading and summarizing the dataset directly (not generating its own random sample — that would be Track B work).
  • The figure does not appear (Track A). The chunk silently errored on an unknown column name. Check colnames(data). Quotation marks in aes() are a classic Week-8 carryover bug — see Lab 6’s troubleshooting section.
  • The interpretation paragraph contradicts the rendered output. Re-render, then re-read. If your prose and the PDF disagree, fix the prose to match the PDF, or fix the analysis or simulation to actually produce the result your prose describes, then re-render and re-read again.
  • The PDF is suddenly enormous. Either an oversized embedded image or an accidentally printed long vector. Summarize instead of printing.

The render-twice habit is the load-bearing finishing move for both tracks. The PDF the LMS grader will see is the second render, not the first.

AI in the R Project

AI assistance is appropriate for the same kinds of help you have used across Weeks 7–9: R syntax lookup, debugging a chunk error, explaining what a chunk does, rephrasing prose, catching a typo, suggesting a label for a plot axis, reading a Quarto or LaTeX render warning. See the AI use guidelines for the full pattern.

What AI cannot do in the R Project:

  • AI cannot read your rendered evidence for you. Track A: what the rendered analysis shows about your dataset is what you see in the rendered PDF, not what an assistant predicts the analysis would show. Track B: what the simulation under your seed produces is what is in the rendered PDF, verified by re-running with the same seed yourself.
  • AI cannot decide whether your interpretation is honest. Honesty about what a small dataset or a small simulation can and cannot support is your call.
  • AI fabricates dataset URLs and citations. Anything you cite is your responsibility to confirm.

The Verification line of your AI Use Note is the load-bearing one for the project. It must name what you actually checked — not “I used Claude.” If you did not use AI, say so on the Tool line; that is a complete AI Use Note.

Where the project folder lives

The project’s portfolio location is math-software-portfolio/r-project/, alongside latex-project/ from Week 6 and the weekly hw01/–hw09/ folders you built in Weeks 1–9. The folder is the project’s permanent home — you finish it in place and submit it from there.

The exact project material — including the prompt and the submission details — lives in the course LMS.

Looking ahead

Module B closes at the end of Week 10. Next week, Module D opens with the AI module (Weeks 11–12). Those two weeks carry the two non-droppable weekly AI assignments — the AI module’s parallel to the LaTeX Project (Module A) and the R Project (Module B). The Week 13 portfolio/workflow conference is where the finished r-project/ folder is reviewed as part of the portfolio next to the finished latex-project/ folder and your weekly artifacts.

Submit the R Project in the state you want to read again at the end of the term.