Today I went to a course by Edward Tufte, the Yale professor and
information organizer extraordinaire, at the Grand Hyatt in downtown San Francisco. Despite the prohibitive price tag ($380 for six hours), I was really excited: he has beautiful work, and you get fine hard-cover editions of most of his publications besides. I knew some of his work, but had never read the books. They are gorgeous and worth every penny.
Tufte is lean and small and academic, with pleated slacks and wire-rimmed glasses. The only jarring things were the massive rings on his fingers (six: four on his right hand, two on his left) and his occasional bursts of irritation at the current administration—otherwise, the whole day turned out to be a plain ol’ college lecture, where we calmly got out our books and turned to this-or-that-a-page and patiently took notes on pads provided by the hotel. Odd sidenote for the techie bunch: I turned check about three times, but
never saw a laptop open in the crowd. One person unfortunate enough to be checking his phone after lunch was duly chastened: “Don’t look at your cell phone right now—we’re talking about
Galileo. Geez.”
The man himself, signing a book
The lecture started off with a nice visual representation of one of Chopin’s lullabies, while Tufte gave an equally nice overview of the levels of information being described: time (past, present, future), spacial representation of tone, color, and so on. Thereafter my notes usually devolved into quotes, which reflect his general themes of data quality, credibility, and simplicity.
When you start to strip detail from your presentation, you lose credibility.
To clarify, add detail.
Document your stuff, and provide reasons to believe.
You’re trying to evoke a content response, not a design review.
Tufte had a passionate aside about “chart junk” and how you should question the intent, nay, the
integrity of the organization who would present data with bright blue gradients. On a practical level, he mentioned twice that cartography is a good reference point when trying to display data that overlays other figures. In fact he recommended stealing and copying whenever possible:
Look for excellent conventional displays, in most cases.
The idea in analytical design is—don’t be original, do it right. Steal from good examples. Find a good metaphor.
For those who would decry large and complicated tables as an efficient way of displaying data, he mentioned stock charts and sports statistics. I love this; so often in web design I hear that things need to be dumbed down in order to be digested by the masses. I’ve often wondered what the real benefit is: less customer service?—is that the goal of a site?
Nielsen notes users will put up with a lot of irritation before they’ll leave a site, and
Norman adds that users will actually blame themselves when things are poorly designed. Clearly life should be better, and design can improve everyday irritations substantially. Nonetheless Tufte puts an emphasis on taste and quality over accessibility, and this is refreshing.
It seems there is a misplaced egalitarianism about design that I’ve noticed in many a dull meeting: something like, “Ten-point Hoelfer Text may be all well and good for you design types, but fourteen-point Arial will be understood by those people who [somehow] cannot read or understand anything else. Nevermind that people visiting the website/reading the article/looking at the sign will have overcome many a better-designed obstacle to get where they are, looking at our material; when it comes to us,
concessions must be made.” Tufte does not hold with such nonsense. About page three, my notes summed up the general theme “…He continually reminds you that your audience is not stupid, and is in fact as capable as you are.” But here are a few direct quotes:
Think of your audience as busy, not stupid.
People aren’t going to get any more confused than you are.
(And how to avoid confusion in general? “Annotate everything,” and a reminder that in most cases, with tabular data, “This is a solved problem.” Additionally his love for sidenotes matches my own, and makes me wish I’d found a reasonable sidenote solution for this blog with CSS. If anybody knows how to do it without absolute positioning, let me know.)
The last few paragraphs are not to indicate that Tufte prefers design to usability. In fact quite the opposite. There is no mercy for those who would “put lipstick on a pig,” that is, dress up otherwise useless information. Tufte actually prefers annotations to a lot of usually unobjectionable cartographical conventions: “…codes, legends, and other fooling-around are impediments to learning.” This is an uncharacteristically website-friendly point of view, where the static position of a legend leads to a lot of useless scrolling. Frankly, it’s clear that
legends can be useful, when they are useful and easy to memorize; Tufte’s examples of their failings were all print, and involved awkward page breaks. But there’s a lot to be said for describing visual verbs, like arrows: just adding a bit of descriptive text can make up for any number of awkward map keys.
After lunch, I was hoping for more in-depth practical examples, and maybe some guidelines, about the visual problems Tufte had described earlier. Since I’d arrived ten minutes before the course started, I hadn’t realized (unlike this attendee) that a lot of what he was saying aloud was quoting verbatim from his books. I did, however, already knew about sparklines, and was disappointed to find that was all Tufted had slated to talk about for the next hour. Sparklines are tiny little fever graphs, the height and width of regular words, designed to be used inline with text as a tidy visual of whatever data you’re interested in, and they are super neat. Their practicality is obvious and the idea is really easy to grasp. In short, they are so fabulously intuitive they do not warrant an hour’s worth of explanation.
Tufte ended the day with (a) a long analysis of why Powerpoint, and more generally shallow outlines, totally suck (
and can lead to death), (b) how to glean the correct information, or at least know what you
should be looking for in a meeting; and (c) how to shorten up your presentations.
Along with most people in the audience, I already hate Powerpoint, not to mention any designer has a burning desire to destroy any program that only allows nudges in inch increments and has a native resolution of 180 x 200px—so Tufte’s explanations added fuel to the fire, but did not really change anybody’s mind. As for meetings, I’ll duly write what he prescribed, but I’ve never been in charge of a dull presentation, and aim to
never be in charge of one. Here are the tips:
- Only show full-sized, useful pages and videos during presentations. Otherwise:
- Write up an executive summary, about two hundred words or so, and
- Then outline the rest in a nice tabloid-sized paper, printed on both sides, which costs very little, and can hold an amazing sum of data, which is useful because
- The person you’re presenting to is quite busy, and can read a lot faster than you can talk, especially with those ridiculous Powerpoint slides*
*I added that last bit.
More importantly, he added a few key things to remember if you are ever an Important Person getting information from a Powerpoint or other badly organized presentation:
- What is their story?
- Can you believe them? What’s their credibility?
- What is the domain specification? (In other words, to what, exactly, is this relevant?)
- What do you really need to know? (And this last question should get you out of the room.)
A lot of the second half of the lecture was so transparent that I can’t wholly recommend the course to interested designers. It’s worth getting the books, but once you have them, the data overlaps too much to justify the price tag. Having said all that, Tufte is interesting and charming, and a good speaker. I’ll leave you with a few last quotes on creating interesting information graphics:
Analytical design is content-based, which in turn leads to elegant but transparent architectures.
The idea is to create interesting comparisons. The question is always, “Compared to what?”
Compared to
what? That’s always a good question.
—Timoni
