Training opportunities

Notes and advice from the Edward Tufte classroom

It was truly a pleasure to attend Edward Tufte’s one day conference today in Durham, NC. I was able to get two books signed (one for me, and one for you!), some questions answered, but most importantly, I learned an incredible amount about interface design, data visualization and graphic standards. Included below are his tips and comments throughout the lecture. I didn’t use quotes on anything just in case I misplaced a word here or there, but this advice came straight from him. Enjoy!

Remember, he will be doing eight other conferences around the U.S., so make sure to attend one in your area!

Also, recruit friends to II and track it on our wiki to win an autographed version of “The Visual Display of Quantitative Information.” Deadline is this Friday at 5pm!

Tufte Advice

On good practices …

• Focus on causality
• Rely on multiple sources and levels of data
• Don’t pre-specify the type of approach and/or method
• Process driven vs. Content driven
• Annotate linking lines (links as verbs)

On data design …

• Find first-rate examples and follow them
• Build 4-5 templates and stick to them
• Gills Sans is always a great typeface for tables
• Always attempt a super graphic
• Find intellectual models
• Get credibility, include external sources

On principles of analytical design …

• Show comparisons, contrasts, and differences
• Show causality
• Show multivariate data (3+ variables)
• Integrate multiple mediums; do not segregate by media type
• Document everything and be transparent about it
• Serious presentations stand or fall depending on the quality, relevance, and integrity of their content

Tufte quotes:

About design …

There is no such thing as information overload, just bad design. If something is cluttered and/or confusing, fix your design.

High resolution designs are genuinely interactive.

Negative space is optical clutter. The way to reduce 1 + 1 = 3 (two objects plus negative space) is to reduce the contrast between figure and ground.

Never do lowest common denominator designs and/or presentations.

Small multiples are easy on the user’s eyes. Your goal is to maximize the content reasoning and minimize the format understanding time.

Local optimizing = global pessimizing (When everything is highlighted, nothing is highlighted)

Detail specificity gives you credibility.

Thoughtful designs can eliminate half of legends.

The principles of analytical design are the same as those of analytical thinking.

About graphics …

If a digram is worth 1,000 words, then the user should spend just as long examining a graphic as he/she would take reading several paragraphs of information.

Maps are the heart and soul of this field. They are almost entirely content, they use color, typography, and are efficiently designed.

Our eyes see the equivalent of two ethernets and process three dimensional data all the time … why are we worried about information overload with two dimensional charts?

Every graph you see should give reasons for you to believe. This should be followed by a universal, parenthetical caveat (“This graph is accurate until alternative explanation or better evidence comes along.”)

What you want when thinking about data display is an open mind, but not an empty head.

Why should we allow the plumbing to determine how we display our information?

Annotation is at the heart of explaining things. Annotate everything with linking lines with ties to the element.

Never order things alphabetically unless you are designing the phone book. Order by performance; why are we measuring this stuff? Ordering by performance allows the user to learn something in the scanning process, unlike alphabetical lists.

About Powerpoint …

Powerpoint is extremely narcissistic with an attitude. It’s a cognitive style with a contempt for the information, therefore a very poor presentation medium. With powerpoint, the name of the presentation method has wiped out the content (It’s my powerpoint presentation vs It’s my cancer presentation).

The average Powerpoint presentation has 12 numbers on it, which is 5-10% of an average sports section or weather page.

We’ve got to get out of this authoritarian style of presentation. It disrespects the audience and their cognitive skills and abstracts from their interests.

We want to display information adjacent in space rather than sequential in time (use of charts in presentations versus slides).

Presentations should be handout-driven and not slide-driven. Your new presentation software is Word, not Powerpoint.

The frequent error is to underestimate your audience, which leads to dumbing-down the presentation.

Begin every presentation with a super graphic to get your audience thinking.

On interface design …

The design principles are irrelevant to the mode of display.

Alan Cooper: “No matter how beautiful your interface is, it would be better if there was less of it.”

Information should become the interface.

No icons! Icons are useless because they still need words underneath as a description.

You don’t need to attract people to your website (via design), they’ve already arrived (via content).

Failure to display content is swiftly punished. The back and close buttons are designers’ two biggest competitors.

Content-oriented interface design with 90% of space devoted to content is ideal … soon to be 94%.

© 2012 Innovative Interactivity (II).
Download II's Security Policy (PDF)