Beautiful Evidence, the latest book by Edward Tufte, lives up to its name. The book itself is strikingly beautiful, starting with the dust jacket, continuing through the page design and layout and, most importantly, the many reproductions of excellent, mediocre and abysmal examples of graphical presentation of data. Indeed, this book would be worth its purchase price merely for the illustrations, which Tufte informs us "come from 14 centuries, 16 countries (Italy and France, especially), 3 planets, and the innumerable stars." The choice of such varied illustrative manner is no accident: Tufte states in the introduction that he believes that "the principles of analytical design are universal" and this book is sort of an extended discourse on that theme. Tufte states further that he believes that "making an evidence presentation is a moral act as well as an intellectual activity" which explains in part why he reacts so strongly to poor or misleading graphical presentations.
Edward Tufte is most noted today for his work in the field of informational graphics. His previous books include The Visual Display of Quantitative Information (1983; 2nd ed. 2001), Envisioning Information (1990), and Visual Explanations (1997). Further information, including a moderated forum and a number of articles by and about Tufte, are available on his website: http://www.edwardtufte.com.
But it is difficult to imagine that anyone with a serious interest in the graphical presentation of information would not already be familiar with Tufte's work. If they do exist, such readers may become confused by what sometimes appears to be the random presentation of topics, and by the far-reaching generalizations of the author. Beautiful Evidence is a very individual book, and anyone who buys it expecting a standard textbook on graphic design will be disappointed.
The apparent disorganization is deliberate: one of the evils Tufte is fighting is a mechanical approach toward data presentation, which encouraged by cookbook methods presented in many textbooks and by many software programs commonly used to produce statistical graphics. Rest assured, there are in fact unifying principles behind the many examples, although the connections between the various examples may not always be immediately obvious. Readers who want an overview of Tufte's principles can start by reading the chapter "The Fundamental Principles of Analytical Design." Otherwise, I recommend starting anywhere and stopping to carefully study those examples which are most personally appealing. For instance, I particularly enjoy Tufte's suggested revisions of published graphics, such as those from Markus Bloch's Ichthyologie and from Carl Sagan's The Dragons of Eden.
Tufte does have a tendency to express himself in extreme terms, but taking issue with particular statements or principles seems to me to be beside the point. We can argue forever about whether the principles of analytical design are really universal, but it is not necessary to settle that issue in order to benefit from the material presented in Beautiful Evidence. The occasionally harsh tone of Tufte's criticism is offset by his considerable wit. My favorite example is from his commentary on Galileo's Sidereus Nuncius, which included engraved images of the moon disproving the contemporary belief, based on Aristotle, that the surfaces of all celestial bodies were smooth and regular:
This evidence-free fancy had become religious doctrine before being demolished by 10 pages of visual observations reported in Sidereus Nuncius. Galileo later questioned why religious faith need depend on the presence or absence of craters on the moon or, indeed, other astronomical observations. (p. 98)
Readers who have read Tufte's previous books and articles on graphic design will find that some of this material is familiar, particularly the chapter on "The Fundamental Principles of Analytical Design" which is based on Charles Joseph Minard's data-map of the manpower losses of the French army during their 1812 invasion of France, and the chapter "The Cognitive Style of PowerPoint: Pitching Out Corrupts Within," much of which was previously published in pamphlet form.
Sarah Boslaugh (firstname.lastname@example.org ) is a Senior Statistical Data Analyst in the Department of Pediatrics at the Washington University School of Medicine in St. Louis, MO. She wrote An Intermediate Guide to SPSS Programming: Using Syntax for Data Management with Sage Publications in 2005 and is currently writing Secondary Data Sources for Public Health: A Practical Guide for Cambridge University Press. She is also Editor-in-Chief of The Encyclopedia of Epidemiology which will be published by Sage in 2007.