Edward Tufte, Professor Emeritus at Yale, Author of Seven Books including "The Visual Display of Quantitative Information"

A couple of weeks ago, Edward Tufte was in Boston giving a seminar. I've been pretty busy lately, but have heard good things about Tufte's ideas on presentations so I made the time to attend one of his courses. It ran from 10-4:30pm with a 1 hour lunch break. Along with the session, I got copies of three of Tufte's books, a poster of Minard's Napoleon's Death March, and a pamphlet on Powerpoint. It cost me $180 because I was able to get the student discount, but normally it is $320, which is still pretty cheap. Overall it was a good value.

Tufte is an interesting character indeed. His teaching style is unique. He obviously has a deep understanding of the material and tries to teach it in such a way to give the student an appreciation for presenting information in a useful manner not just an understanding of how to do it. He also doesn't hesitate to speak his mind even when his opinion is negative (especially regarding Powerpoint).

I've posted my notes below. Tufte references his books throughout the session and constantly showed visual examples (including really old first print editions of books by Euclid, Galileo, and Newton). My notes are just text so it doesn't have quite the same effect.

Notes:

One way to think about your presentation: what is the rate of information transfer (or resolution)

Tufte reviewed Euclid's book on Geometry

He referred to "escaping flat lands" many times; the thought is that we have to describe a three dimensional word in two dimensions which is difficult. In some cases, such as Euclid's book (Tufte has also done it in his book), he included foldouts that help the reader visual certain points.

One way to escape flat land is to use a model

He described 8 principles of analytical design:
1. Show comparisons
2. Show causality, process, dynamics, and mechanisms
3. Show multi-variant examples (more than 1 or 2 variables)
4. Completely integrate word, number, and image; it is all evidence
5. Document everything and tell people about it
6. Presos stand or fall on the quality, relevance and integrity of the content; At least "do no harm" to the content; Powerpoint does not meet this minimum bar
7. Try to show information adjacent in space rather than stacked in time.
8. Use small multiples. He showed Galileo's example figure of sun spots – showed one big sun and 37 smaller versions to illustrate that sun spots move; Great for comparisons; Easy on the viewer; Has inherent credibility; Deflects suspicions of cherry-picking

Your evidence (or cherry-picked evidence) is an important issue to consumers

It is important to put your name on your work to show personal responsibility

At the beginning of a preso, you should cover the problem, relevance and solution

Tufte analyzed magicians because they try to hide information. He wanted to under their principles and then attempt to do the reverse (to unhide information)

He likes wall charts for managing projects, not a single computer screen (because it has such low resolution)

On a presentation, make remarks as small as is readable

Tufte does not like letter codes used within figures or graphs; You should be able to look at a figure and understand it without needing to constantly refer to a legend in a different location.

It is ok for the viewers of your preso to look at different parts of the slide than what you are currently discussing

If it looks cluttered on a computer screen, fix your design. A computer screen can't hold too much data – its capacity isn't big enough

To clarify, add detail – don't remove detail

Need simple underlying designs with rich, complex information

The viewers time should be spent rolling around in all of the content

Simple design and intense content

How to display financial data:

Consider what is the thinking task that this preso is suppose to help with?

Graphs don't have to show the 0 point

Show more data horizontally than vertically

Confirming a previous known detail helps credibility

Use annotations to help with causal explanations: "Annotate everything"

Steal from the best – use excellent conventional designs

Look a the Wall St Journal or NY Times – they have it right – 3 million sold everyday

"Talent imitates and genius steals" –TS Elliot

Sparklines (more available on Tufte's website):

You should try to be approximately right rather than exactly wrong

Intense content, simple design

There are only two industries that refer to their customers as users: Software and Illegal Drugs

For authoring sparklines, he recommends Adobe Illustrator, SAS or some stat package, and Quark Express or some publishing/layout tool. Currently it isn't cheap.

Web Design:

Design mimics bureaucracy

Two most embarrassing words in web design: Skip Intro; This forces users to become interface designers of a site

Mimic one of the good existing sites (he recommends Google News)

Use links as a last resort. Scrolling is ok.

Need to consider the computer's more precious resource: screen real estate

There is a contest for screen space: coming in first is usually OS imperialism; then browers, apps, etc. take up the other portions

Remember that navigation is not content

90% of every screen should be content; Most are 15-30%

All you need is a navbar at the top (not the side or bottom). The rest should be information

He doesn't think there is / needs to be a "web design" field – should be more like web editing: editing content, not designing it

Cognitive Style of Powerpoint:

Tufte reviewed his analysis of the Columbia accident

Powerpoint is medieval with its preoccupation with hierarchy

If you see test data, you should ask yourself it is truly independent or is the author not showing us something (ie Cherry-picking)

Powerpoint is not a serious method of presentation

There should be no bullets, no indentions, no logos

Use full screen images and full sentences

Provide a hardcopy sheet of data to your audience

Your presentation display technology should be Word (ie, Word should replace Powerpoint)

Don't announce you are using something other than Powerpoint – Don't attract attention to the presentation method

Presentation Guidelines:

Show up early to your own presentation

Tell them up front what the problem is, the relevance to them and the solution

Use paragraph form with sentences, no bullets

Never apologize in your opening (eg, Sorry I'm late, Sorry we had to stay late, etc)

Always give the audience a piece of paper (tech report style)

Think about what your audience reads

Only show full screen images

Make the presentation a content experience not a design or presentation experience

Alienate your audience on the merits of your content not a ghastly throw away joke

Think the best about your audience (respect them) don't treat them like idiots

Get up close and be appropriately demonstrable; don't just clutch the podium

Finish early

Before you give a presentation make sure you Practice, Practice, Practice and get the best content you can