Student Materials Repository for CS 8630 - Data Visualization
Every assignment has three parts: Read & Write (framework-grounded prose from your own reading), Build (a code deliverable), and AI Instructions (what AI use is required, permitted, and prohibited). All assignments live in your course repo; every graded artifact is committed with its AI-LOG.md entries in the same commits.
The assignment list is tentative after assignment 3, and may be adjusted based on student interests, contemporary data visualization questions, or emerging tech.
| # | Week due | Unit | Title | AI Tier |
|---|---|---|---|---|
| 1 | 2 | Perception (Cleveland, Franconeri) | The Perceptual Evidence Brief | 2 |
| 2 | 4 | Grammar of Graphics (Wickham, Wilkinson) | One Dataset, Ten Specifications | 2 + one Tier-1 exercise |
| 3 | 5 | Semiology (Bertin) | Generate–Critique–Repair | 1 |
| 4 | 6 | Multidimensional & Hierarchical | Beyond Two Dimensions | 2 + one Tier-1 exercise |
| 5 | 7 | Graph & Network | Node-Link vs. Matrix | 2 |
| 6 | 9 | Geovisualization | The Lying Map | 1 |
| 7 | 10 | Text Visualization | What the Word Cloud Hides | 2 + AI-annotator audit |
| 8 | 12 | Persuasion & Misinformation | The Honest Persuader | 1 |
(Week 11's assessment unit is examined through the in-class Misleading-Visualization Turing Test lab, whose artifact feeds Assignment 8.)
The prose is yours. In every tier, AI may not draft your essays, memos, or critiques. Permitted: asking AI to critique a draft you wrote, to find counterarguments, or to check your use of a technical term — all logged. Prohibited: generating prose you then edit, and using AI summaries in place of doing the reading. Reading engagement is verified live: each unit opens with cold-call discussion of the readings, and your Week 15 defense can reach back to any assignment. Prose that outruns its author is obvious in ten seconds of conversation.
Code is collaborative, logged, and owned. Tier 2 means AI can scaffold, debug, and suggest — and every accept/reject decision that touches an encoding choice goes in the log with its perceptual justification. You are examinable on every line you submit.
Tier 1 means the AI's output is course material. Where an assignment directs you to prompt a model and analyze the result, submit the raw output unmodified. Curating flattering outputs defeats the assignment and is treated as data fabrication.
Accuracy is yours. Unchanged from prior years: an AI-introduced error in your submission is your error.