Amanda Cox — legal full-text readings

Focus: Cox-bylined or Cox-spoken material available as full public text (interviews, conference write-ups with substantial quotation). “Above all else show the data” is Edward Tufte’s line—not attributed to Cox here.

NYT Upshot interactives and the Reader Center profile repeatedly timed out on fetch; those are flagged NOT read in full (possible soft paywall / heavy JS). No Sci-Hub or pirate archives used.

---

1. Interview with Amanda Cox — Graphics Editor at the New York Times (Simply Statistics)

• - URL: https://simplystatistics.org/posts/2012-06-01-interview-with-amanda-cox-graphics-editor-at-the-new/
• - Date: June 1, 2012
• - Read in full: YES — Cox’s own Q&A answers

Notes (what SHE argues in this interview)

1. 1. She declines the “statistician” label: doesn’t treat uncertainty formally enough; good days the work is “journalism.”
2. 2. Path: restless first-year UW stats → applied for random things including NYT graphics internship.
3. 3. News graphics ≠ scientific audience graphs: “feels like homework” is a negative reaction.
4. 4. Practical differences: don’t assume readers already care; kill jargon; graphics should often stand alone; typography/design = hierarchy and clarity, not cuteness.
5. 5. Prototypes in R (base points/lines/text); sometimes clean PDFs in Illustrator; web interactivity changes the R stage into proof-of-concept.
6. 6. “Poor-man’s interactivity”: pagedown through a PDF of charts from a loop.
7. 7. Topic choice: news value and interestingness; ideally both.
8. 8. Data availability shapes scope (e.g., county vs precinct election maps in real time).
9. 9. Turnaround: “how long do we have?”—projects expand to fill time (hours to weeks); desk proud of one-day breaking work (often less data-heavy).
10. 10. Annotation layer is critical: annotate example points so readers learn forms by example; words should highlight the relevant pattern or an expert interpretation—not “here is some data.”
11. 11. Likes sliders (suggest beginning–middle–end even when not time) and meaningful motion that exploits web vs print.

---

2. The NYT’s Amanda Cox on Winning the Internet (OpenNews / Source report of OpenVis Conf keynote)

• - URL: https://source.opennews.org/articles/nyts-amanda-cox-wins-internet/
• - Date: May 17, 2013
• - Author of page: Erin Kissane (conference report)
• - Read in full: YES — detailed secondhand write-up of Cox’s OpenVis opening keynote (not a word-for-word transcript)

Notes (arguments attributed to Cox in THIS report)

1. 1. Question: what makes some Times graphics “better,” and which kinds of better matter?
2. 2. Social shares ≈ batting average: imperfect quality metric, not uncorrelated with some quality.
3. 3. Initial guesses for “winning the internet”: surprise; illusory comprehensiveness; goofy comparisons; clear takeaway.
4. 4. Then she manually coded top ~400 graphics on many traits (surprise.reveal, emotional/atmospheric, useful, development.really.hard, takeaway, etc.) and graphed results.
5. 5. Empirical top traits for social success: hard development; breaking-news-adjacent; useful; explicitly emotional/atmospheric; surprise reveal; comprehensive.
6. 6. Difficult topics (war, climate, etc.) did worst; blunt “takeaway” pieces also underperformed as a class.
7. 7. Signaling importance: use scarce resources—print “good page” placement vs web scarce developer hours.
8. 8. Two “wows” (via Jonathan Harris / Ed Ruscha): spectacle wow vs intimate connection wow; prefer “Huh? Wow!” over “Wow! Huh?”
9. 9. Design-as-empathy realization: not mainly whitespace—but helping readers understand and feel.
10. 10. Little templating—“the best journalism isn’t a mad lib”; craft to the data’s unique shape; abstract frameworks OK.
11. 11. Transparent versioning (GitHub) so remote colleagues see visual evolution.
12. 12. Sampling/randomness on web (e.g., county political “wind” map) can show “all at once” differently than print’s forced emphasis.
13. 13. Closing challenge: tools hide things—“What are you skipping because it doesn’t come for free in your tools?” Goal: find what she’s missing because she thinks differently.

---

3. Amanda Cox: Making Illustrations Better with an Annotation Layer (Eyeo 2011 attendee notes)

• - URL: https://www.softwareandart.com/amanda-cox-making-illustrations-better-with-an-annotation-layer/
• - Date: June 27, 2011 (Isabel Wheeler notes from Eyeo)
• - Read in full: YES — attendee notes, not official transcript

Notes (Cox claims as recorded here)

1. 1. Dataviz can learn from illustration: foreground what you care about; leave some detail out; “how many backup singers?”
2. 2. Mythic over-detailed diagrams (every helicopter bolt) aren’t always necessary; essence often enough—but sometimes detail helps.
3. 3. Principle: show a pattern, leave some detail (like reporters’ anecdotes).
4. 4. Examples: Mariano Rivera pitch video (pattern of indistinguishable spin at decision point); tables surfacing what maps hide; comment-sentiment grid; ski-jump video telling viewers when to look at arms/hips.
5. 5. Add guidance to prevent misinterpretation; avoid scavenger-hunt interfaces.
6. 6. Online free-for-all (“here’s data, tweet if you find something”) abdicates responsibility.
7. 7. Need a proper annotation layer—don’t say “here’s an interface, browse forever.”
8. 8. Sketch on paper through multiple forms until something works.

---

4. Isolating the elements of compelling graphic design (Nieman Lab)

• - URL: https://www.niemanlab.org/2013/05/isolating-the-elements-of-compelling-graphic-design/
• - Date: May 20, 2013
• - Author: Caroline O’Donovan (link-blog summary of the OpenNews piece)
• - Read in full: YES — short secondary summary; largely duplicates #2

Notes

1. 1. Reframes Cox’s question: what response do we want readers to have?
2. 2. Repeats the six social-success traits and scarce-resource signaling argument.
3. 3. Little independent Cox text beyond what OpenNews already reported—treat as pointer, not primary source.

---

5. How a statistics major became graphics editor at The New York Times (Nieman Lab)

• - URL: https://www.niemanlab.org/2012/06/how-a-statistics-major-became-graphics-editor-at-the-new-york-times/
• - Date: June 11, 2012
• - Author: Joshua Benton (link to Simply Statistics)
• - Read in full: YES — very short; quotes annotation-layer passage from #1

Notes

1. 1. Points readers to the Simply Statistics interview.
2. 2. Highlights “poor man’s interactivity” and the annotation-layer quote.
3. 3. Not an independent Cox essay.

---

6. Amanda Cox Promoted to Data Editor (NYT Company press note)

• - URL: https://www.nytco.com/press/amanda-cox-promoted-to-data-editor/
• - Date: Jan 15, 2019
• - Authors: Dean Baquet & Joe Kahn (institutional)
• - Read in full: YES — not Cox-authored

Notes (institutional claims about her role/work)

1. 1. New “data editor” role to connect CAR, Upshot, Interactive News, Graphics.
2. 2. Career arc: 2004 graphics intern → helped launch Upshot (2014) → Upshot editor (2016).
3. 3. Desk credited with dialect quiz, you-draw charts, election “needle,” and visual evidence on racism, gender inequality, opioids.
4. 4. Advises masthead on polling/forecasting methodology.
5. 5. Useful for biography/context; not a design manifesto in her voice.

---

Not read in full / blocked

---

Attribution caution (per brief)

• - Do not credit Cox with Tufte’s “above all else show the data.”
• - Cox’s distinctive public teaching emphases in the readable corpus are: annotation layer, show a pattern / leave some detail, empathy/“right wow”, tool-blind-spot self-critique, and empirical coding of what “wins” attention (OpenVis).

---

Search-loop note

Searches for “Amanda Cox Upshot,” “OpenVis,” “annotation layer,” and related phrases returned the same OpenNews, Simply Statistics, Eyeo-notes, and NYT profile URLs. Stopped when the next searches were duplicates.