Edward Tufte Review

from the

New York Times

UP FROM FLATLAND

By Phil Patton

In 1570, the first English translation of Euclid’s “Elements of Geometry” was published. On the page that discusses the pyramid, the printer pasted in a small paper model that can be folded upright from the page. One copy of the book found its way into the library of Ben Jonson. Today it belongs to Edward R. Tufte, a 55-year-old professor of political science and statistics at Yale University. At home in Chester, Conn., Tufte holds the book in his lap and carefully unfolds the thick, darkened paper pyramid from the page. “Only Ben and I have done this,” he says with a mischievous grin.

For Tufte, who reproduced the standing pyramid in his book “Envisioning Information,” published in 1990, the pyramid is a miniature monument to the dream of escaping the two-dimensional surface. “Escaping this flatland is the essential task of envisioning information,” he writes. “For all the interesting worlds (physical, biological, imaginary, human) that we seek to understand are inevitably and happily multivariate in nature. Not flatlands.”

Tufte (pronounced TUFF-tee) has become a prophet of “information design,” the design of graphs, charts, diagrams and other visual elements. With more and more facts and statistics available, information design has become a focus in increasing attention in the news media. Graphs and charts can replace thousands of words or numbers, economically summarizing and explaining things on the page or screen. The last few years have seen the arrival – in newspapers and magazines, on television and at business presentations, from USA Today to CNN – of a new generation of innovative information designers. Richard Saul Wurman, a designer and publisher, breaks cities down into neighborhood maps, visually signifying restaurants, shops and so forth, in his Access travel guides. Nigel Holmes, in his book “Pictorial Maps,” outlines the ideas behind the cartoonlike “infographics” he creates for Time magazine. Tufte himself, with his focus on fidelity to statistics and historical principles, has consulted for CBS, NBC, Newsweek, The New York Times and the Bureau of the Census. IBM has hired him to review the design of computer interfaces and manuals.

By Tufte’s calculation, there are between 900 billion and two trillion statistical graphics printed annually worldwide. “Clutter and confusion are failures of design, not attributes of information,” he declares. “There’s no such thing as information overload.” Too much data, perhaps, but not too much information. His scrutiny and analysis have been directed at everything from weather maps to the inscriptions on the Vietnam Veterans Memorial, from diagrams of the life cycle of the Japanese beetle to maps of the distribution of air pollutants in the Los Angeles basin and cholera in 19^th-century London. Some are as witty as the chart showing the food provisions of an ocean liner, including a 4,340-pound chicken; others as dramatic as a chart of Tufte’s own design tracking the booster-rocket problem that killed the Challenger astronauts.

“I don’t want people to ever see the same way again,” he says.

Working on a research project at the National Center for Supercomputing Applications, at the University of Illinois in Urbana-Champaign, Tufte is moving from flatland to “movieland,” attempting to arrive at principles for making clearer the moving models – generated by computers – of complex scientific theories, from the interaction of atoms to the theoretical physicist Stephen Hawking’s theory of the universe.

One recent morning at the center, Tufte, in torn jeans and loafers, bounds downstairs into a darkened basement room where a half-dozen huge computer screens glow. He has worked most of the night and is still excited. “Wow,” he says, “I haven’t worked that late in a computer lab since I was an undergraduate.” (He went to Stanford.) He has come here to help redesign a classic example of computer visualization, “A Study of a Numerically Modeled Severe Storm,” better known simply as “Storm.”

For two and a half hours on April 3, 1964, a thunderstorm raged across Oklahoma and Texas, spinning off a tornado that killed seven people, injured 111 and did some $15 million in damage. Now that thunderstorm -- or at least its numerical ghost – surges across a computer screen, compressed by a supercomputer into a video model, six minutes long. Such models may eventually help predict the course and severity of storms and tornados.

With the help of computers named Klimt, Courbet and Mondrian, Tufte is doing here in computerland the same things he does in flatland, simplifying the design but adding layers of detail; lightening the grid across which the storm moves, adding a “tripod” scale showing distance and compass direction. “We want to increase the resolution of the space,” he says. “It’s moving, but it’s all information.”

Tufte suggests putting a “movie strip” of storm images at the bottom of the screen to function as a visual clock, to show the viewer where he is in the sequence. It uses one of his favorite ideas – small multiples. “That’s great!” he exclaims as it comes on the screen. “Wonderful. It’s just what Scheiner, Galileo’s rival, did with sunspots in 1626.”

Tufte has a copy of Christopher Scheiner’s book at home, and one of Galileo’s. Along with the Euclid, for which he paid $19,000 at Sotheby’s, they are part of what he calls his “Museum of Cognitive Art.” He draws both inspiration and examples from such books, written, he says, when “word and drawing were as one” and the image was “just another sentence element” in the argument. For all his work with computers, his ideas center around a ringing affirmation of the power of the book.

His own books are designed so that visual examples merge seamlessly with text. In “The Visual Display of Quantitative Information” (1983) and its sequel, “Envisioning Information,” Tufte has traced the history of the discipline, showcased the best examples and formulated principles to guide future designers. The books are a collection of solutions in information design, drawing on examples from the work of unheralded pioneers like the Englishman William Playfair, whose 1786 chart traced the size of the British government debt, and Charles Joseph Minard, A French engineer whose 1869 graph-map tracking Napoleon’s 1812-13 march to and from Moscow Tufte feels “may well be the best statistical graphic ever drawn.” Tufte argues for simplifying graphics in order to make them more complex, removing unnecessary elements in order to introduce new, useful ones. He speaks of the “smallest effective difference” – the lightest line or shade or simplest shape that will do the representational job. Ideally, the design should disappear in favor of the information.

That his examples are also visually rich, colorful and intriguing in form is an indirect result. “I view esthetics as a fortunate byproduct of the display of information,” he says. There is something Nabokovian about his approach, with mixes the scientific and the artistic, and about his language, in which the technical mingles with the metaphorical. “Ideas about information design have a wonderful scope and reach,” he says. “Such ideas are indifferent to disciplinary distinctions, indifferent to whether it is science or art.”

The principles of good design are also “indifferent to time, indifferent to space.” Tufte’s examples of good design include a Japanese spy’s copy of a Japanese railroad schedule, reproduced in “Envisioning,” and a courtroom chart given to him by John Gotti’s lawyers, Bruce Cutler and Susan Kellman, which helped lead to the reputed mobster’s acquittal in 1987. The chart listed the Government witnesses against Gotti and their own criminal activities. “Though only 37 percent of all the possible combinations are marked,” Tufte writes, “they dominate the spreadsheet grid.” Where crime and witness intersect, the mark is a powerful X. Together, the 69 bold X’s impugn the Government witnesses with all the power of an old-fashioned warning on a poison bottle.

Bad examples are easier to find than good ones. Tufte fights what he calls “chartjunk” – which he defines as “content-free decoration” – and quantifies the “lie factor” of bad graphics, like those that use dollar figures without adjusting for inflation. He castigates drawings of oil barrels that exaggerate the increase in petroleum prices (the volume of the barrels increase in three dimensions, but the price in just one) and singles out as “chartoons” graphics like one that traces the price of diamonds on the bent leg of a chorus girl, using her fishnet hose as a grid.

Some practicing graphic designers see Tufte’s standards as overly academic and rigorous. Time’s Nigel Holmes, creator of the diamonds graph, was understandably irked when Tufte criticized it. Holmes admits his work has sometimes been exaggerated, but feels that Tufte, in his insistence on absolute mathematical fidelity, remains trapped in “the world of commerce,” with its need to grab an audience. But even Holmes admires and admits to having been inspired by Tufte’s books.

Tufte’s rule is that any graph should have a true scale – numbers should be represented accurately by the dimensions of ink on the page, as in a map scale. He wants to add data so the viewer can personalize the image, read it his or her own way. “Detail lends credibility,” he proclaims. For the train schedules Tufte and his wife, the graphic designer Inge Druckrey, conceived for New Jersey Transit, he adapted a structure devised in 1846 by Charles Ybry, but he added high-resolution aerial photographs marked with the course of the rails. “Look,” the image invites, “you may be able to find your own house.” To Tufte’s delight, many riders reveled in looking for their houses, but to his disappointment most found the schedule no easier to read than the previous 32-page all-type version.

To demonstrate that information design can be a life and death matter, Tufte pulls out a chart prepared by Morton-Thiokol, the maker of the space shuttle’s solid-fuel rocket boosters, relating damage to the boosters to the temperature at launch time. The design of the chart – with launches arranged according to date not temperature – hides the fact that, as Tufte puts it, “There is all the information necessary here to have saved the Challenger astronauts.” Tufte redraws the chart to make the point clear – damage had most often occurred before at low temperatures, and no launch day had been nearly as cold as the 31 degrees predicted for Challenger.

Tufte’s care with visuals extends to his own books. “Control of every detail of publishing,” he says, “is an extension of my means of expression.” He co-designed “The Visual Display” with Howard Gralla and designed “Envisioning Information” himself, using a computer. He supervised both books through the printing process, checking proofs with a magnifying glass. When the pop-up pyramid in “Envisioning” began showing up with wrinkles, he put the original Euclid in the back of his Toyota, drove to his bookbinder in Charlestown, Mass., and showed him how the original was done.

“ This paper is 420 years old and hasn’t puckered yet.”

Finding that no academic or commercial publisher could produce “The Visual Display” to his exacting standards, Tufte published it himself after taking out a $100,000 second mortgage on his house to finance it. Under the imprint “Graphic Press,” the books have now sold 750,000 copies. Graphics Press has made several million dollars and become a legend in the publishing world. ID magazine gave “Envisioning” its coveted designation of best graphic design of the year; the designer Milton Glaser, one of the judges, called it “one of the best books on graphic design ever published.”

Says Tufte, “I can’t tell you how much of a dreamlike quality there is about the work.” That ideal intellectual landscape seems to have its physical counterpart in the eight acres and house in Connecticut where Tufte and Druckrey live with their three golden retrievers. Around a swimming pool is a fence Tufte designed, which, like a trio of his own sculptures dotting the lawn, is made up of wooden beams arrayed in a pattern of diagonals reminiscent of the New Jersey train schedule design.

In his studio, filled with piles of books and papers, Tufte is working on his latest book, to be called “Visual Explanations.” One of the key ideas of the new book is that of the “visual confection” – a term he has coined to describe “an assembly of several visual events,” which are “brought together and juxtaposed on the still flatland of paper.” An example of this is the classic illustration of how to perform a magic trick, which combines movements over a period of time on a single page. Tufte also cites Eadweard Muybridge’s 19^th-century photographs of the gait of a horse as an example of a visual confection.

Along with the book will come a videotape illustrating his recent endeavors in movieland, including the redesign of “Storm.” To Tufte, the danger in the movieland of computer-generated graphics lies in what he calls “dequantification”; the numbers get lost in a fascination for shapes and effects. Scales are dropped, meaningless colors are added and “molecules dock to the tune of space music.” Such graphics are inspired less by information design than by television, which Tufte detests. “It’s paced to the human voice, which is slow. It’s violent.” And, Tufte’s ultimate dismissal: “It’s very low resolution.”

By afternoon in the computer room, new effects are being added to Storm.” Klimt displays alternate typefaces for the scale. “It’s a detail, but the details accumulate,” Tufte explains. “It’s design by aggregate pickiness.” On Courbet, alternate grids are tried out. On Mondrian, Tufte watches as the storm unfolds over and over again, growing from a bead of mercury into something like a slug, then an anvil, then a delirious version of Dumbo the elephant. The storm slides above the grid, slips in front of the tripod scale. Tufte is delighted. “That’s a sweet touch!” he exclaims, “that revealing and concealing. The scale is there but it’s harmless. You know right away it’s supporting information, not primary.”

Flying home that evening, his plane passes over Chicago. It is a breathtakingly clear night, at rush hour. The grid of the city glows. Lending scale are the Hancock and Sears towers, shifting in relationship to each other on the horizon, alternately concealing and revealing each other as the plane banks. Below, the solid cherry glow of taillights in a traffic jam constitutes a virtual graph of driver frustration. “Wow,” Tufte says as he leans toward the window for a better look. “Look at that resolution.”

It’s all information, after all.