in Uncategorized

Tufte’s Beautiful Evidence, Chapter 2

Earlier this month, I wrote an at-times-savage review of the introduction and first chapter of Edward Tufte’s new book, Beautiful Evidence. I write that critically because I hold Tufte to very high expectations — if only because he considers himself a critic of the highest order.

Chapter 2 is markedly different from Chapter 1 in that it is actually and believably brilliant. Titled “Sparklines: Intense, Simple Word-Sized Graphics,” it provides detailed presentation of a new form of information visualization that can be embedded in the flow of text.

Something fascinating about this chapter is how I knew everything to expect, because Tufte wrote this chapter very much in public. Over two years ago I wrote about an early discussion of sparklines. And Tufte’s discussion boards are filled with commentary and examples dating back to mid-2004.

Sparklines are a brilliant and seemingly inevitable innovation in the display of data with text. The degree to which it can enhance understanding, particularly of anything statistical, is enormous.

I do find some of the illustrations needlessly obtuse (the human genome, 3D scatterplot). And the inclusion of another Durer drawing seems less about intense, simple word-sized graphics than about showing off his appreciation of fine art.

But my primary frustration is Tufte’s continuing dismissal of computer screens in favor of high-quality printed paper. Sparklines, as I wrote back in 2004, seem ideal for computer displays — with their ability for color, animation, and, most important, interaction, just as much, if not more, information can be presented on screen than on print. Unfortunately, when Tufte sees a computer, he seems to see a static medium, and so doesn’t recognize the power. Hans Rosling’s displays at TED demonstrate that animation over a single screen can provide a far more instructive experience than seeing the information displayed over small multiples.