Month: June 2007

As with his books, Edward Tufte suggests that your designs should be self-revealing. That is, they should practice what they preach. For example, his books on visual design are well-designed.

Not only that, but his presentation on presenting information, “Presenting Data and Information,” was very well presented. I was very impressed at the end of the day when his last point was “Finish Early,” and he said it three minutes before we were supposed to be done. What a great way to teach…

I highly recommend this conference by Edward Tufte, no matter what arena you work in. Check the E.T. web site for details on when/where to attend.

I attended the seminar on May 24th, 2007 in Crystal City, VA. I found the best way to take away his message was to encode small ideas as sentences.

Notes

Display of data should be driven by its content, not by what you are “good” at.

Use all the levels that it takes to explain the meaning of the content (not just one level).

Details matter: annotate relationships between items (it helps your credibility with the audience).

Model of a good diagram is a good map; maps generally don’t have non-essential information.

Don’t put boxes around things—instead, use that room to display information.

Provide “reasons to believe” such as a credibility statement. An example is a disclaimer such as, “This is the best we know until better evidence comes along.” Always leave the door open and be open to improvements.

Rich supergraphics supersede just generic bullet lists (E.T. doesn’t like the bullet “reveal” technique commonly used in PowerPoint presentations).

Aim for rich content and very simple design. This will maximize content reasoning time and minimize design “figuring out” time (decoding, fooling around, and figuring out time for people viewing your work).

Tables will outperform graphics for small datasets (less than 300 numbers). Example of sports pages and financial data (mutual fund page): these should be our models (you won’t find zebra strips or fancy icons there).

“Don’t get it original, get it right”: use excellent but conventional templates.

Research is more important and appropriate than creative work. Know your content!

Design standard: at least do no harm to the content.

Follow the metaphor of teaching and educating people through the content and design; use elite newspapers as examples: NY Times Thursday science articles for example.

The principles of analytical design are based on the principles of analytical thinking (and also therefore tied to nature’s laws)

The goal of information design is to aid thinking

• Example: if you want to compare things, show comparisons directly in your design.
• Example of Mihard’s “Napoleon’s March to Moscow” graphic: Mihard didn’t make the graphic because he was good at it (drawing, engineering). He was trying to make a strong antiwar statement. The word “Napoleon” doesn’t appear on the map anywhere; it’s about the soldiers. He cared about the content so much that the result is the best it could be. The design is obsessed by content.

Show important information side-by-side (avoid the flip, flip of back and forth movement to compare things visually). For example, the medical bill chart with annotated sidenotes is a good template for reports (explains the data with side-by-side notes).

Good quote: “If you look after truth and goodness, beauty looks after herself.”

Sparklines: based on the idea that graphics are no longer important; they are the same as letters and numbers.

“Nature” magazine: best statistical graphics available (also see “Science”).

Content matters most of all.