A Tufte Library
To me, the most important books on presenting data graphics are by Edward Tufte. People who create charts as part of their job should keep one or more of them close by, and regularly reread them.
The Visual Display of Quantitative Information, 2nd ed.
Graphics Press, 2001 (orig. 1983) • ISBN: 0961392142
This is the most essential Tufte book, published over 20 years ago but so far ahead of its time it looks absolutely contemporary. (Compare it with some of the other data graphics books from the mid-1980s and you’ll see). Tufte here introduces his recurring themes: maximizing the data-ink ratio, stripping away unnecessary furniture and “chartjunk”, showing all the data, and graying out what’s less important. There’s elegant discussion of how readers perceive area changes, why we shouldn’t think of data as boring stuff that needs livening up, why pie charts suck, and when tables are better than graphs. Some of my favorite examples from here are Minard’s famous chart of the retreat from Moscow, the step-by-step erasure of needless ink from a bar chart, and the use of one of his own early graphs as an example of poor design. And the whole book is beautifully typeset and produced, with restrained use of color and plenty of white space.
Envisioning Information
Graphics Press, 1990 • 0961392118
Similar in subject and tone to the first book, though there’s more discussion of mapmaking. The most important concept discussed here is small multiples, the fruitful idea that six small graphs instead of one large one can add a new layer of information and reveal large patterns in the data. I’ve personally found this the most enlightening technique to share with newcomers to information design, and it’s often the best way to break out of the prison of an unsatisfactory design solution. Tufte also discusses how to use color effectively and with restraint, and how to successively reveal a process in a linked sequence of diagrams (such as dance notation or calligraphy). I particularly liked his coverage of tabular data, from train and bus timetables to the Vietnam War Memorial. Another fine book.
Visual Explanations
Graphics Press, 1997• 0961392126
This book wanders further out of strict chart-and-graph territory to cover depictions of processes, exemplified by the diagramming of magic tricks; if we understand how misdirection works, we can turn it around to draw the reader’s attention to what really matters. There are nice examples of the practical consequences of information graphics, discussing their role in the Challenger disaster and John Snow’s investigation of the London cholera epidemic. The most valuable concept to me was the notion of the smallest effective difference, with a useful discussion of just how little emphasis is needed to make a point clear in a graphic, and the power of using gray shades and faint lines. (There’s a fun genealogy of rock ’n’ roll too, illustrating parallelism as a graphical tool.) Lovely, but a less immediately practical book than the two preceding it.
Tufte’s newest, Beautiful Evidence, has just been published, and a review will appear here as soon as my copy arrives.
Comments
going to a tufte day-course is great too. the books are included in the registration price and students usually get a hefty discount. i went for $135 (student plus group discounts), and considering i learned more in 1 day than i learned in 4 years of college, i think it's money well spent.
Posted by: jessica | August 10, 2006 11:45 AM
Personally the one thing I remember best (and I can't remember in which of the 3 books I saw it) is Tufte's redesign of how to represent the musical score (in western standard annotation). (Tufte suggests fading out the obvious information and emphasizing the informative parts such as the actual notes etc.)
Obviously the current design will not be changed within the next 50 years or so. However, I totally agree with Tufte that it would make sense to change it... which has given me a lot to think about.
Posted by: elias | December 6, 2006 1:07 AM
I wrote this some months ago in Stephen Few's discussion board:
(...) What I find interesting in Tufte is that he builds a theory of information visualization based (among some other things) on some very strong aesthetic principles from minimalism (Mies van der Rohe). The result is very coherent and it gives you a strict and normative way of looking at graphs ("maximize data/ink ratio"; "maximize data density"; "no chart junk", etc.). But design is, by its very nature, a subjective answer. You can't always back it with scientific evidence. You can't anticipate all the logical consequences of your design theory and some of them will collide with scientific findings, sooner or later.(...)
I also see myself as a Tufte fan but are there any other options? Tufte gives us something unique: an holistic theory and metrics to evaluate the results (the full monty...). No one else does that, only some random guidelines. I suspect that Tufte's theory is too positivist and works better inside a simplified reality. I am glad we have him around, but I think we need an alternative view (just in case...).
He is Marx, not Leonardo.
Posted by: Jorge Camoes | February 26, 2007 5:19 PM