28 June 2012
Tufte tackles reckless information design
Beautiful EvidenceBy Edward Tufte<br>Graphics Press, USD52<br>
Edward Tufte is design’s champion of reason. His principles of scientific rationalism are old school, but his expressions of analytic design are sophisticated, smart and painfully relevant. Eight years in the making, Beautiful Evidence, his fourth book, makes a soberly brilliant case for renewing responsibility in design.
Now Professor Emeritus at Yale University, Tufte has taught courses in statistical evidence, analytical design and political economy. Over the years, he has examined a wide variety of visual explanations (from Galileo’s notebooks to charts in Nature magazine) and posted his findings on his website, spurring discussion among his many faithful readers. His prose can be dry, awkward and jargon-filled (‘What are the content-reasoning tasks that this display is supposed to help with?’). His prose can be forgiven, however, because Beautiful Evidence showcases Tufte’s wide-ranging eclecticism while grounding itself with a genuine and serious respect for truth.
Adhering to principles of observation over superstition and evidence over sensationalism, Tufte is a friend of science and statistics, a rational idealist who presumes that truth is a shared social value. Where a newspaper would print a photograph of Saturn, Tufte would add an image of Earth for size comparison. Where a nightly newscast would display a spiking line graph in shocking red, Tufte would add values and a reference to a mean over time. Where bar graphs abound and pie charts predominate, Tufte would wipe the messy slate clean and bring proportion and clarity to our media-driven infoglut.
Divided into seven main chapters (the last 24 pages are devoted to sculpture, including Tufte’s own), the book proceeds by a case-study method, examining specific designs for their virtues and vices, and extracting general rules. He appreciates the efficient and elegant integration of text and image as much in a field guide to birds or a report for detecting concealed weapons as in Leonardo da Vinci’s notebooks or Galileo’s The Starry Messenger (1610).
In the chapter on ‘The Fundamental Principles of Analytical Design’, Tufte derives his rules for good design from rules for good thinking, and he identifies expressions of these rules in Charles Joseph Minard’s data map of casualties during the French invasion of Russia in 1812. Tufte’s first principle is to show comparisons, and he then explains how Minard’s map ‘makes several vivid comparisons’, most notably the difference between the size of the army at the start of the campaign (422,000) and, six months later, at its finish (10,000). Tufte uses the remarkable map to illustrate his other principles, including to show causality, to show multiple variables, to integrate words and images, and to provide documentation of sources and authority. These commonsense rules are controversial for being so widely ignored in print and broadcast media, in online news sources, in medical journals, textbooks and legal documents. In every chapter, he takes pains to point out specific instances of design negligence, from nasa’s PowerPoint presentations on the 2003 Columbia disaster to a chart showing the relationship between brain mass and body mass for 26 animals.
Tufte walks his talk by improving specific designs. In the second chapter, he discusses sparklines, which are ‘small, high-resolution graphics’ or ‘datawords’. His first example is a patient’s glucose level. As a single number, the glucose level has little meaning. Tufte adds context (a data line of the previous 80 readings) and a grey band on top of the data line to show the range of normal readings. In this way, patients can quickly see the range of their current level compared with their history relative to the normal limits. Tufte then seeks to improve the design, and thus the meaning, of sparklines for currency-exchange rates, mutual-fund performance, and baseball win / loss records. Their designs vary depending on context and content, but the principles carry forward into the next application.
Tufte’s design improvements are so sensible they sell themselves. His sparklines pack a lot of information into a small clean package. His redesigned charts clarify and contextualise. He even describes the tools and capabilities of the perfect PowerPoint killer. Why aren’t more information designers doing what Tufte is doing? We defer too often to the slick tropes of pseudo-science, aided by Microsoft’s PowerPoint (see chapter 7) and abetted by the nightly news.
As an academic and an independent thinker (as well as a self-publisher), Tufte contributes a needed corrective to reckless information design. We depend on evidence to persuade us of an understanding, and we use this understanding to guide our judgements about belief and behavior. Therefore, we need to bring a rigour and a respect to the graphic depiction of evidence of all kinds. Beautiful Evidence should be required reading for precisely all those in media, government and business who think they don’t need it.