Edward Tufte's One Day Course: A Review
Last Monday, I got to attend Edward Tufte’s one-day course. I was looking forward to a day of interesting examples, ideas, and discussions, but was disappointed by the amount of rambling and largely historical examples, with little connection to real, current visualization (or presentation) work.
The Setting
The course took place in the large ballroom of the Westin Seattle, which was set up for around 500–600 people. Maybe it was naive to assume a more intimate setting, but I had imagined around 100 people there. There was, consequently, no interaction with the audience of any sort, other than people lining up to get their books autographed before the course started or during lunch break.
As part of the course, you get his four books in a little cardboard box with a handle. With the box, you are handed a sheet of paper with a reading assignment: one or two chapters from each book in the first hour. Unfortunately, he does not actually make use of that reading in his presentation, presumably because he knows that only a fraction of attendees actually read everything they’re supposed to before he gets started.
The Good
It’s no secret that Tufte likes paper. His sparkline technique is meant to be a high-resolution, high-density data display that communicates a lot of data in the same space as a number (or maybe a few numbers). His relentless pounding on PowerPoint is also mostly focused on the fact that slides tend to break information up into small chunks, rather than laying it all out so it can be taken in.
His suggestion is to hand people the information, in printed form, at the beginning of the meeting and give them 10 minutes to read. That way, everybody can get a sense of what is there, and skip the parts that are not relevant. This makes a lot of sense, though it does assume a certain type of presentation. This will not work for presentations where a certain amount of theatrics is actually important, like a product presentation or even some strategy discussions. For a data-heavy presentation, his method is undoubtedly a good one.
There is also something to be said about the density and size of a sheet of paper. You can fit a lot of information onto a piece of printed page, much more than a typical projection screen or slide (recent high-resolution screens are slowly getting there). However, to make that truly useful, you have to spend a lot more time laying out the information in a meaningful way. Done well, this can be incredibly effective, however.
Tufte also has amassed a huge wealth of historical examples of data visualization, some going back many hundreds of years. The presentation in his books is also incredibly well structured and designed, and he is a good speaker. Some parts of it, such as the beginning, were perhaps a bit overly dramatic, but it was never dull.
Where he is also certainly right is his motto of Do whatever it takes! to get the job done. The point here being that you don’t want to pre-specify the tool, but pick it based on requirements. His rant on how academics tend to work the other way around had a lot of truth in it, though it was of questionable relevance to the audience.
The Bad
Tufte likes to ramble. A lot. I’m not opposed to a good rant every now and then, quite the opposite. But everything in moderation! It seems that he spent half of the entire course time on wild tangents and rants about things, most of which he really does not know all that much about.
One particular example was his long half-joking dismissal of big data. This being Seattle, he undoubtedly had people from Microsoft, Amazon, Google, and probably a dozen big-data startups in the audience. I doubt that they came to hear him dismiss big data analysis as mostly self-correlation. I don’t think he’s entirely wrong about that, but he didn’t exactly offer anything useful, either. He claimed that the scientists of the 16th and 17th century had been the original big data people, because they had to turn their observations into theories.
Some of his other rants included his current sculpture work and how he is now more interested in 3D these days, how people are getting unnecessary medical test because ‘the money is in the false positives,’ proprietary vs. open technology (which hilariously ended with him praising PDF), how administrators are taking over universities, Apple, and of course PowerPoint.
Tufte’s scholarship of old visualization examples is interesting, but unfortunately he is also stuck in the past. He showed off an original printing of one of Galilei’s books, as well as the first translation of Euclid’s work into English. What was the point of that? Blinding us with impressive artifacts?
He had some good points about how Galilei presented some of his observations, like the sun spots he saw and that kept moving in a way consistent with a rotating sun. That made for an interesting history lesson, no doubt, but what can we learn from that?
The Ugly
The rambling is one thing, but the 15-minute sales pitch was quite another. Books! Posters! Yes, Tufte told us to buy his books, his posters, and even his mother’s book on grammar, the only non-ET book Graphics Press sells. You really don’t expect that when you’re paying $380 for the day.
He also seemed very worried about people realizing how important he was. He listed the things he had done, including not only his recent presidential appointment, but also the importance of sparklines, etc. For the latter, he kept citing the fact that Google had returned more results for the query “sparkline” than “Steve Jobs” – as of July 2011.
One of the videos he showed was of his wavefields idea, which he describes as an extension of sparklines. What it is, though, is a bizarre claim about HD movies and data representation that is either a prank (as one poster on his forum also noted) or just shows a complete failure to grasp human perception.
Throughout the course, Tufte seemed out of touch, both with his audience and with the real world. At least three times, he said something along the lines of “We as real scientists …” to an audience that was hardly a group of scientists. His recommendation of tools also included LaTeX, which may still be the academic choice for writing papers with lots of math, and which have to follow strict formatting guidelines, but hardly a practical suggestion for most business users.
Conclusion
My advice? Buy his books. Read them. They’re good. Just realize that you’re getting a historical perspective on data visualization, not the cutting edge. Understand that Tufte’s ideas are a good starting point, not a religion. There are many things that Tufte doesn’t know, including pretty much anything related to visual perception and cognition, recent work (less than 30 years old), and interaction.
Once you’re done with Tufte’s books, read Stephen Few’s or Colin Ware’s or something else that’s recent. But by all means, skip the course.
Posted by Robert Kosara on August 5, 2012. Filed under criticism.