Summary of Edward Tufte's Visualization Course
ddgarcia@cs.berkeley.edu
Overview 
I attended a one-day course titled Envisioning Data and Information
taught by Edward Tufte in the Fall of 1994
and I thought I'd highlight the main topics presented. If you'd like to find out more about these issues, please see me in person.
Basic Problem
Multi-variate data
Everything interesting we want to display is a multi-variate (multi-variable) problem and all we can use is a 2-D space
Resolution
How can we communicate more information per unit time than previously?
Small Multiples
There is an inherent problem when displaying similar images
(e.g. frames from an animation) on separate pages, forcing the reader
to flip back and forth to compare the nth frame to the n+mth frame
known as the one damned thing after another
problem. It is solved by displaying many small
multiples at once. Small multiples are small, thumbnail-sized
representations of the images displayed all at once, which allows the
reader to immediately, and in parallel, compare the inter-frame
differences.
Smallest Effective Difference
Whenever you make a mark (label, line, etc.), make it as small
as possible, but as small as possible to still be clear. This
enhances the resolution.
Causality (Answering the question: "Compared to What?")
Displays should show causality of the various
parameters. The correlation of the presence of the cause -> presence
of the effect and absence of the cause -> absence of the effect should
be investigatable.
Context
Data should be put in proper context. Example:
"I just got a raise of $1000 this year"
tells us nothing. How does that compare to what you were making last year?
How does that compare to raises in the past? How does your % raise compare
to others in the company? How does it compare to others at other companies
doing your same job?
References
Tufte, Edward R. The Visual Display of Quantitative Information, Graphics Press, Cheshire, CT 1983.
Tufte, Edward R. Envisioning Information, Graphics Press, Cheshire, CT 1991.
WWW Maven: Dan Garcia (ddgarcia@cs.berkeley.edu) 
Send me feedback
