Unordered todo list. Pointers: (PLAN) = MODULE-DOCS-PLAN.md · (BRIEF) = starter brief in Workspace/Briefs/ · (MINE) = analyzed in _lecture_experiment/lecture-mining-plan.md (slide refs, figure candidates) · (SUMM) = AI reading summary ready in Workspace/Re-Papering/summaries-*.md · (DRAFT) = draft page exists, needs review.
Order by importance for class.
- Look at 0718 - it has a "next steps for mike"
From 0718
- Review encodings
- Review data abstraction
- Review charts-are-encodings
- Try looking at briefs
Useful at start of class
- Tutorial 0
- Tutorial 1 Split
- Critique Tutorial (pull in more book stuff, ready for class)
- Re-papering rant
- Favorite visualization
Building Blocks pages
- Abstraction definition from 765-25
- Data Abstraction tutorial
- Task Abstraction
- Repapering Task Cubes
- Repapering Munzner Tasks
- Repapering Problem Space
- Encodings Tutorial
- Charts are encodings
- Graphical Perception Tutorial (C&M repapering might be separate)
- Low-level tasks
Hidden in Problem-Space/todo.md
- Repapering Project
- get the "rant" page finished
- connect the rant page to the papers page
- connect the rant page to the initial examples
- Finish the Problem space re-papering
- Pie Chart Experiments - update and finish
- How to Vis (or get rid of)
- Comparative Rivers and Mountains (best vis ever)
- Encodings tutorial — review draft. All open edits are now consolidated in
Workspace/Briefs/04-encodings-review.md(a retro-brief; these pages predate the brief format). Two findings from a deck nobody had extracted (4-2, "Encodings Practice", 9/28): your own slide defines Pareto mapping and flags it "(mike's term ?)" with three benefits better than the snack's wording — that pending edit is answered; and Cairo TT ch.5 vs Munzner ch.5 contradict each other on hue's rank, both required in Module 2, which the "Which Channel Is Best?" section should own. Still needs: genai credit line, parallel-coords "best of a bad lot" rewording, image provenance check; publish together with charts-are-encodings (cross-linked) (DRAFT, PLAN, BRIEF) —content/tutorials/encodings/index.md - Charts Are Just Encodings in Disguise — review draft; decide merge-vs-differentiate with the old "Chart Typologies" idea. New evidence (
Workspace/Briefs/02-vis-basics.md): the real argument is in lecture 2-M slides 5–38 (the LLM chart-naming experiment + the unemployment walkthrough), not 4-M — a tutorial's worth, which would swamp this 682-word snack. The compressed version here is the thin one and should probably become a pointer (DRAFT) —content/snacks/charts-are-encodings/index.md - Data Abstraction — two competing drafts, pick one. (a) long tutorial
content/tutorials/data-abstraction/index.md; (b) cheat sheetcontent/tutorials/data-abstraction-cheat-sheet/index.md(2026-09-06, derived from (a); motivation deferred to Munzner, encodings link dropped, new categorical-⊂-nominal section). Both draft=true. Then: verify the NOIR table, verify the categorical claim (see brief §7.2), decide if NOIR splits into its own snack, wants figures. BRIEF + report inWorkspace/Briefs/. - Pie Chart Experiment — long-standing draft snack; the perception/evaluation pages will want to link it (DRAFT) —
content/snacks/pie-chart-experiment/index.md - My Favorite Visualization: rivers & mountains comparison — draft critique (DRAFT) —
content/snacks/260520-compare-rivers/index.md - Re-Papering rant — draft explains the whole re-papering section; publish when first re-paperings do (DRAFT) —
content/rants/repapering/index.md - Problem Space paper page — replace the NotebookLM notes with a real page (DRAFT, SUMM) —
content/papers/problem-space/index.md
- Tutorial 0: the prime directive — "make visualizations to address tasks." Working framing: what is vis? — and why the answer tells you how to do it and how to learn it. (The run-on-ness of that sentence is exactly Tutorial 1's disease; Tutorial 0 is the cure: one idea, short.) Stub exists at tutorials/0-how-to-vis (DRAFT stub) —
content/tutorials/0-how-to-vis/index.md - Split Tutorial 1 — 6,700 words → some goes into Tutorial 0; keep a tightened core (definition, effectiveness); spin out pieces below; best done AFTER other tutorials exist to absorb material (PLAN, BRIEF: tutorial-1-revision) —
content/tutorials/1-what-is-vis/index.md- Why Visualize? — brief written (
Workspace/Briefs/02-why-vis.md): the four whys as the skeleton, "what are your alternatives?" is the slide nothing on the site covers, four-authors summary is the closer. Checked: Tutorial 1 poses this question in its "Now What?" list and never answers it, so redundancy risk is low — this is a real page, not Tutorial-1 leftovers. Needs §5–8 + the Tutorial 0 collision decision (BRIEF, MINE) - Is a stop sign a visualization? — definition-edges snack; the brief argues it may belong in Why Visualize instead (definition judged by usefulness, not correctness) (BRIEF, MINE)
- Why Visualize? — brief written (
- Most important principles — "if you read only one page" snack; relation to Tutorial 0 needs deciding — may be the same page
- "Vis Basics" / What Different Charts Make Easy to See — 2-M slides 5–38: LLM chart-naming experiment, gallery critique, unemployment-by-state walkthrough. Destination undecided on purpose: own page, the body of Tutorial 3 (1,061 words — it has the thesis and one example; this is the missing practice), or stay a lecture. The fork is whether the LLM experiment gets published — as a lecture it's re-runnable yearly and improves; as a page it spoils a flagship exercise and dates. §1–4 done (BRIEF: vis-basics)
- "Start here" prose in
content/tutorials/_index.md— MIKE writes this. No longer blocked: the structure is settled (2026-08-22). The collection is onetutorials/section — core sequence 1–4 (weight1–4) reads in order, unordered deeper tail atweight30+, Tableau class-support between. What's needed is the framing paragraph plus the ordered "start here" list; a tag can't carry order or "why this one next," so it's prose. Currently one sentence long —content/tutorials/_index.md - Give Tutorial 0 a
weightwhen it's written (no weight today, so it sorts last in the collection — presumably wants 0) —content/tutorials/0-how-to-vis/index.md
- Task Abstraction — highest leverage; "task first" has no home page (PLAN, BRIEF)
- Too Much Stuff (scale) — strategies + throwing-information-away-wisely; connects to comparison paper page (PLAN, BRIEF; Munzner ch. 13/14 SUMM)
- Evaluation — nested model + "Convince [ ] that your [ ] is [ ]" (PLAN, BRIEF; Munzner ch. 4 + North + Cairo TT2 SUMM)
- Implementation: choosing tools — timeless space-of-tools, expressiveness vs. ease (PLAN, BRIEF)
- Learning from Tufte (carefully) / How to Read Tufte — snack; needs tone calibration (PLAN, BRIEF; Tufte VDQI ch. 1 + Cairo FA3 SUMM)
- Encodings Practice: taking designs apart and putting them together — 4-W slides 4–19 + 4-2 slides 3–68. The compose half (2 numbers → scatterplot → parallel coords → +categorical → 3D) and the decompose half (Playfair, Minard, then Cairo's authors problem with "You can do better!"). Slide 60's "Lessons" is the thesis. Answers
MODULE-DOCS-PLANitem 6: the 2-numbers exercise is a tutorial's worth, not a snack, so the encodings page's "Try It" box becomes a pointer. Watch: Dominik Moritz attribution on the decompositions (BRIEF: 04-encodings-practice) - Tableau is the theory, embodied — 4-W slides 20–29; promote off "someday"? It's the only genuinely new + small week-4 doc. Not about Tableau: it's the argument that the encodings framework is real enough that a commercial product runs on it ("blue = dimension, green = measure"; "get your data into the right form so the visualization it makes is the one you want"). Keep it UI-free so it doesn't rot (BRIEF: 04-tableau-embodied)
- Perception for Vis: the fast and the slow — pop-out, attention, "design to use the fast mechanisms" (MINE; Munzner ch. 5/6 + Cairo FA5/FA6 SUMM)
- Why Vis Works (cognition) — Card/Mackinlay/Shneiderman amplify-cognition material; overlaps Why Visualize — decide ownership (MINE; CMS summary pending)
- Vis & Statistical Thinking — descriptive vs. generalization, multiple comparisons, model-vis (MINE; Leek&Peng/Shmueli/Zgraggen summaries pending)
- Color crash course — physics→perception→systems→use; LAB is why theory matters (MINE; Munzner ch. 10 SUMM)
- Too Many Dimensions — glyphs, parallel coords, SPLOMs, scan/subset for dimensions (MINE; Munzner ch. 13 SUMM)
- Embeddings & Dimensionality Reduction — PCA to UMAP; "is it meaningful?"; timely beyond class (MINE)
- Graphs Are Not (Just) Node-Link Diagrams — matrices, reordering, tasks; links route-maps critique (MINE; Munzner ch. 9 SUMM)
- How Node-Link Layout Actually Works — heuristics→objectives→force-directed (MINE)
- Interaction (and its costs) — builds around existing interaction-costs paper page (MINE; Munzner ch. 11/12 SUMM)
- The Third Dimension (a taste) — "we sense 2D, we infer 3D"; depth cues (MINE; Munzner ch. 6/8 SUMM)
- Fields, contours, volumes — MIKE'S CALL whether the site needs it (MINE; Munzner ch. 8 SUMM)
- The Talk on Talks (presentations) — The List + its origin story; job-talk history section is Mike's call (MINE)
- Design School in a Day — contrast/repetition/hierarchy; links design books resource (MINE)
- Uncertainty: what it is before how to draw it — uncertainty vs. distributions; "uncertainty vis == model vis?" (MINE)
- Find the red dot (pop-out demo)
- The eye is not a camera (retina bandwidth, computer-systems framing)
- Can you see the average? (visual proxies — verify whose studies)
- Stream graph aggregation gotcha (from a real student question)
- Why the rainbow colormap misleads ("use order for order!")
- Metamers, or how displays fake out your eyes
- How many colors can you actually use?
- Parallel coordinates: order really matters
- Radar charts vs. star coordinates
- What does distance mean in an embedding?
- Scagnostics: finding the interesting scatterplots
- Reorder the matrix, reveal the structure (Bertin tie-in)
- Venn diagrams don't scale (UpSet does)
- Animation: use with caution ("How you will die")
- Why does the light come from above?
- The Curse of Knowledge
- What do slides actually do?
- The weather forecast problem (conveying a prediction)
- Where you put the bins changes the story (spin-off of Too Much Stuff) (PLAN)
- Overdraw (spin-off of Too Much Stuff) (PLAN)
- Expressiveness vs. ease of use (spin-off of Implementation) (PLAN)
- NOIR levels of measurement (possible split from Data Abstraction)
- Composing designs: the 2-numbers exercise (only if split from encodings page) (PLAN)
- Tableau is the theory, embodied (someday)
- Bates Cairn (concept)
- How to critique experiments (concept — pairs with pie-chart-experiment and Evaluation)
- Cleveland & McGill (both papers, probably one page) (SUMM)
- Shneiderman, The Eyes Have It (SUMM)
- Amar/Eagan/Stasko low-level tasks (SUMM)
- Mackinlay 1986 (SUMM)
- Bertini, Beyond Precision (SUMM — maybe resource page rather than re-papering)
- North, Toward Measuring Visualization Insight (SUMM)
- Card/Mackinlay/Shneiderman intro; Leek & Peng; Shmueli; Zgraggen (summaries pending — see
Workspace/Re-Papering/status.md) - Possible: averages/visual-proxies papers, scagnostics, embedding-guidance papers (surfaced by MINE; verify which)
- Terrible Treemap (started - in critiques) https://www.insidehighered.com/news/students/academics/2025/08/29/survey-college-students-views-ai#
- The Unreasonable Effectiveness of Pie Charts
- Are tables visualizations? (Yes. and it is useful to think of them that way. And to think of other vis as tables (bar charts, matrices))
- NYT map critique
- Yeping's Graph critique
- Cairo Lines snack
- Chart Typologies → absorbed into Charts Are Just Encodings in Disguise (pending the merge decision above)