Tuesday, July 24, 2012

Presenting Data and Information: A Day with Edward Tufte

Last Friday I had the pleasure of attending Edward Tufte's lecture entitled "Presenting Data and Information" in downtown Denver.  Tufte is a professor at Yale where he teaches courses on statistical evidence and information design.  He's been called the DaVinci of data and graphical information, and one of his books (The Visual Display of Quantitative Information) is widely considered one of the greatest non-fiction works of the 20th century.  For many data geeks like myself, it is a go-to treatise referred to routinely.

Speaking of those books, all four were included in the $380 cost of the lecture, which made the price tag more palatable since they retail for $40 to $50 each.  And they are gorgeous!  Self-published by Tufte, he practices what he preaches and has produced what are truly heirloom quality collectibles.  Everything about these volumes is beautiful: the paper, type-faces, use of color, and printing quality.

So what about the lecture?  For those already familiar with Tufte's work and presentation philosophy, there was nothing new or earth-shattering offered.  The lecture could have been called "the world according to ET in six hours."  His disdain for Powerpoint is well-known, and he did not disappoint in reinforcing this.  Having said that, I truly enjoyed seeing his well-rehearsed and choreographed lecture, and it was nice to hear what he considered the most important take-aways from his four volumes.

My summary of those take-aways is as follows:
  1. All problems worth talking about are multi-dimensional.  They shouldn't be presented serially or one-dimensionally, which is what Powerpoint forces us to do.
  2. Start the presentation by giving a thorough data-dump on a high-resolution display (i.e. a printed document), and spend the first 15 minutes of the meeting telling people to read it.  For comparison's sake, printed documents have 6 to 7 times the resolution of most computer screens, and still twice the resolution as the latest retina displays.  Your audience isn't stupid.  They routinely read and understand charts and tables with 200 numbers or more in the newspaper, magazines, and internet.
  3. Follow that with a summary of the items you wish to highlight and discuss, and open it up for their questions and discussion.  The whole purpose of visually presenting data is to improve understanding, so that's what the focus of the meeting should be.
  4. You don't need to be a designer.  Find some good non-fiction reporting sources that you like, and model your documents after those.  Just copy the fonts, color-schemes, spacing, etc.  Sources that Tufte repeatedly referred to were the journal of Nature, espn.com, plos.org, the New York Times, and the Washington Post.
  5. Show your sources, and make your data available to anybody who wants it.
Would I pay $380 to see Tufte again?  Probably not.  But this is one of those things that I'm glad I did once.  (For more info, check out Tufte's website.)

No comments:

Post a Comment