| Tom 的个人资料Tom Harpel's Space照片日志列表 | 帮助 |
|
2007/8/16 Seminar Report: Tufte's Presenting Data, Seattle, July 24, 2007Update: See http://www.tandoku.com/tufte for a refined version of these notes.
On July 24, 2007, I attended a seminar titled Presenting Data and Information by Edward Tufte in Seattle. Much to process. Key themes:
· Bullets are a despicable form of withholding information · Do whatever it takes to present the content · Do not let the mode of presentation harm the understanding of the content · Handouts are higher bandwidth than slides · The map is the best content presentation metaphor
Other quick thoughts: · First and second act were strong. · The third act was a little unsteady. What did he have at lunch? · Tony Cuozzo was in attendance. We belly ached and riffed over Ivar’s at lunch as only GSO veterans can. · My Day With GFS presentation on capacity planning fails on almost all of Tufte’s principles of content presentation (below); rework required! I can’t believe 1,300 people have so willingly tolerated my defective presentation. The next round will be stronger, I assure you. · Stop by my office to peruse the books, they are beautiful.
Here are my raw notes from the event, transcribed from paper.
1. Opening a. Chopan and midi-like visualization of piano piece i. Represents the notes, the patterns, past, present and future in one view b. Has us reading from the books like a bible at church service 2. Theme for the day: Do whatever it takes to convey the information 3. Info flow practices a. Focus on causality, you can only control the situation by regulating the cause b. Multiple sources of data c. Attitude: Find whatever data you need to convey the truth d. Diagrams are analyses e. Ignore your own style, strengths, tendencies and comforts 4. Annotation is credibility and texture a. Linking lines are verbs. Annotate linking lines to add texture to the relationship. b. On an org chart, annotate link lines to differentiate types of relationships 5. A cluttered presentation is evidence of bad design, not bad data 6. The eye-brain system is high bandwidth a. System runs on 10MB at 16-bit color depth per eye b. Visual information overload is impossible c. The eye is capable of about 1000x bandwidth of a PC display 7. The map is the chief metaphor when presenting data 8. Provide reasons to believe a. Provide the parenthetical (this is the truth until a better view of the truth comes along) b. Leave the door open to getting trumped by better data; this is progress! 9. High resolution data presentation is genuinely interactive a. With sufficient data presentation resolution, each viewer will engage their own cognitive patterns to process the portion that interests them most b. Gets lots of people thinking about your data 10. Supergraphics a. Aerial photographs b. Cancer maps c. History of pop music d. Texture, interactive, transparent design, rich, luscious content presentation e. Avoid contraptions 11. If your numbers are boring, get better numbers a. It’s about your content b. If your content stinks, then your presentation will stink (duh) c. Boredom is a function of your content, not your presentation 12. Graphic layout rule: 1+1=3 a. Two lines (1+1) creates a third element, the white space between the lines b. Lower the contrast (gray lines instead of black, for example) to reduce impact 13. Bullets a. Profound denial of information b. The slow reveal angers Tufte! 14. If you bring handouts and the audience reads ahead, GOOD! GREAT! Your goal as a presenter is for your audience to engage with your content. 15. Always use annotation as much as possible. 16. Tables a. B.E. pp 176 – slops in table, very cool! b. Resort to graphics only to represent 500 or more data points, otherwise use a table c. Performance data always belongs in a table 17. Tufte’s all purpose advice a. Steal from the best i. “Talent imitates, genius steals” T.S. Eliot ii. Get out of the amateur design business iii. Employ taste in choosing what to steal iv. The New York Times, the Wall Street Journal, the journal Nature are great sources v. The NYT has a 30 person data design staff! b. Avoid pitchy, kinky marketing design c. Target documentary presentation of evidence and content d. View design as a research activity, not a creative activity 18. Don’t dumb it down or condescend a. “People have not suddenly become stupid just because they come to your presentation.” – Tufte b. Your audience is smart c. They read the newspaper every day and know how to process graphics and tables 19. Project management: Employ the wall chart with 11 pt font a. Spatiality overcomes sequential presentation of content 20. Break 21. Everything interesting is a multi-variate problem a. 3 or more variables 22. Flatlands: Our task is to boil down multiple dimensions to two for presentation a. Data density and dimensionality go hand in hand b. What’s the bandwidth of your presentation? c. Minimize design admiration time, maximize content messaging time d. Use models when possible to escape Flatland for “real land” e. Build a little model 23. Napoleon’s march a. B.E. pp 124 b. 6 dimensions of data c. Ultimately, an anti-war statement 24. The principles of content presentation a. Show comparisons b. Show causality c. Show multi-variate data d. Integrate data, tables, diagrams, graphics, words, references, qualifications in one presentation or diagram e. Document your sources i. This is what 6 pt type is for! f. A note on these principals i. There is one meta-principle here: Analytical content design and analytical thinking are the same process ii. Thinking = design iii. The components of analytical thought processes are the components of content design and presentation iv. This is about turning fundamental cognitivie process into design process g. Serious presentations depend upon integrity, quality and relevance of the content i. Get better content! ii. Make sure the mode of presentation does not harm the content iii. “Galileo sounds a lot like Feynman” iv. “Both from Brooklyn” h. Do your important comparisons in adjacent space, and not stacked in time i. On one page i. Use small multiples i. Leverage the audiences understanding of a common format ii. Show all of the data j. Aside i. Cherry picking is the greatest threat to truth in content design and presentation ii. The character of evidence should make the point, not the selection of evidence 1. Cherry picking is ultimately fraudulent 25. Sparklines 26. NASA Columbia disaster post-mortem a. Use Powerpoint as a slide projector b. Use handouts, not slides, to convey data 27. Presentation advice a. Preparation i. Practice! In front of a video camera! Brutal! ii. Consider improving the content. b. Beginning the presentation i. Avoid the opening universal joke. Ugh! Ack! ii. Show up early. Get to know people as they trickle in. Some good will come of it. iii. Open with Problem, Relevance to audience, Solution in about 100 or 200 words c. Ending the presentation 28. Tufte has ambitions to sculpt a. Like Judd and Richard Serra b. Cool stuff! c. Does he know Michael Pestel? 引用通告引用此项的网络日志
|
|
|