Wednesday, June 13, 2007

Escaping Flatland...

I almost posted something about Edward Tufte a day or two ago, when Arts & Letters Daily posted a link to the first of two pieces about him. Instead, I hesitated... and I'm glad I did, since the second link is even more Tufte-like. I also figured that the select group that gathers here would already be familiar with Tufte and his revolutionary notions about combining qualitative and quantitative data in ways that allow telling a story.

But then I thought, why not discuss him anyway?

Two short excerpts from the second link:
“We had numbers and words in the same house,’’ Tufte says, as if that explains how he got to where he is. [His father was an engineer, and his mother began a career as a journalist, but later became an English professor.]
In an interview in the mid-’90s with the Computer Literacy Bookshops in San Jose, Tufte explained that to do statistical design, one has to be able to see and to count. He claimed he didn’t see as well as many graphic artists and didn’t count as well as the best statisticians, but he did the combo better than just about anyone. His great insight was to think about graphics not as art or statistical constructs but as stories. He challenged chart-makers to ask the question: what is the story we’re trying to tell?
Both pieces are worth reading... just as Tufte's own site is well worth exploring.

1 comment:

William Timberman said...

Like Buckminster Fuller, Tufte is what the Brits call a one off. His work reminds me again of concrete poetry, poetry which is laid out and printed in such a way that the graphic design of its typography supports or extends its meaning. (Odd, I hadn't thought of concrete poetry in ten years or so, and now it pops into my head twice in a week.)

Purely linguistic communication in the real world is often mixed with other cognitive devices on everything from road signs to computer applications to film. We've adapted to this pretty well. We're happy with Shakespeare on a bare stage, and with nineteenth-century novels, but we also get movies with the partial, disjointed dialogue which mimics ever more closely the fragmentary nature of real-world conversations. (This wasn't always the case, either. I remember the general consternation forty years ago, when Antonioni's Blowup was released. Virtually everyone complained that they couldn't follow what was happening. Show it to an audience today, and no one misses a beat.)

Of course, one may argue that the shortened attention spans which seemingly result from such movies, or from TV commercials and the like ruin us for Shakespeare, but I'm far from convinced that this is the case. We've added other arrows to our quiver, but I'm not so sure that we've discarded the ones which were there before.

Tufte may have an angle on why this is so. As our media change, so does the mix of text and image, or text, numbers and image. As with pure rhetoric, or pure writing, this can be done well or ill, but as long as our educational system teaches best practice, we're not necessarily facing cultural decline because of it. A hundred plus years ago, not everyone wrote as well as Jane Austen or Abraham Lincoln, whatever we may imagine today.

With all that as preamble, this quote from KM's post stuck out for me:

Tufte explained that to do statistical design, one has to be able to see and to count.

I would add one thing. One must also have time. If we're to go from being writers, or statistical analysts to bricoleurs of meaning, we have have time to do some pondering. What was wrong with NASA's charts seems more likely to me to have been the fact that a) they were produced by a committee, and b) they were produced under time pressures, and in an environment of such information overload that no one in charge of compiling the information, or presenting it to the decision makers up the ladder had the leisure to think deeply about how to rank the significance of the information which they were communicating.

This is common in the office cultures that I'm familiar with. Manuals and presentations are hard to compile, and the time pressures are always extreme. Even the most talented at producing them are often reduced to the lowest common denominator. Put the data in, cover your ass, and move on to the next project is the sad rule governing so much of our collective endeavor these days.