Reading Time: 4 minutes

What makes a great infographic? Edward Tufte, in his book Beautiful Evidence (2006), proposes the six basic principles of “analytical design” (he means infographics). Are these the principles that everyone should follow when designing information graphics?


Well, Tufte’s own principles do not always seem to be justified by the examples he provides. For Tufte it is fundamental that infographics should be placed inside of documents, not inside separate application programs. “Segregating information by its mode of production, convenient and profitable for software houses, too often becomes a corrupting metaphor for evidence presentations.”  Here there is a sign of Tufte’s complaint about the use of digital tools to present information – something that led him to write the second part of the book, his famous (or infamous) diatribe against PowerPoint.

Elsewhere, Tufte gives examples of what appear to be excellent integration of text and graphic – but more than once does not attempt to explain what the graphics say. He complains about graphics with arrows, stating “arrows and links are too often non-specific, generic, identical, undifferentiated, and ambiguous”, which may well be true. But he contrasts examples of poor graphics with a double-page illustration from Galileo’s notebooks, pages covered with lines, and Tufte states simply “This 1610 drawing shows carefully articulated links, as detailed annotations describe and differentiates most links. We should do as well.” In my view, we would all benefit from a more detailed explanation about what exactly is taking place in the page illustrated – Tufte gives no explanation at all. It looks beautiful, but what does it mean? What is Galileo trying to say here? We have to take Tufte’s word for it that the links illustrate Galileo’s argument.  

What is Tufte’s real message here? It is a common one, that illustrated publishing went wrong when mechanical tools (and hence by implication digital publishing tools) began to be employed to create graphics. Tufte shows a page from Leonardo da Vinci’s notebooks, as a model of the integration of words and pictures, then describes it in the following way: “this bond between verbal and nonverbal evidence has sometimes come undone in the process of publishing, as the assorted technologies of reproduction and presentation have segregated information by the accident of its mode of production”.  It’s a bit of a leap to blame the failure of integrated text and graphics with the advent of mechanization in publishing. There were plenty of examples of woeful text and graphics in print before digital publishing came along. Perhaps, unfortunately, Tufte means we should all stick to using pen and ink.

Tufte gives as another example of the achievement of integrated graphics that of the famous illustrated book Hypnerotomachia Poliphili, published 1499, which he holds up as “an aesthetic exemplar of printing, layout and typography”. I haven’t read the book, but it certainly looks wonderful as illustrated in Tufte’s book. Yet, as Tufte himself points out, the book fails to integrate the text and the related images. Only 73% of the woodcuts have the related text on the same spread. Some 2.4% of the images have the related text entirely on separate spreads and are therefore impossible to view on the page where the image resides. If this were a book by a modern-day illustrated print publisher, it would be regarded as a failure. Tufte’s suggestion for dealing with this issue? He suggests repeating the image on pages where the text does not fit. This leads to lunacies in Beautiful Evidence such as one graphic repeated five times in thirteen pages (pp 123-24, 126, 128, 135 and 136). It’s a great illustration, but do we need to see it five times? Do we have the space?

Elsewhere, in an assessment of Newton’s Opticks, he complains that in many of the book’s editions the graphics were not integrated with the text: “for Newton’s Opticks, bureaucracies of presentation and mechanical reproduction corrupted the understanding of the content. You’re lucky if they don’t.” According to Tufte the use by the publishing industry of mechanical systems for creating graphics has corrupted integrated publishing.

So what are Tufte’s “fundamental principles of analytical design”? He uses Minard’s celebrated map of Napoleon’s invasion of Russia (1869) to state his “principles for the analysis and presentation of data”:

  1. Show comparison, contrasts, differences
  2. Show causality, mechanism, structure, explanation, systematic structure
  3. Show multivariate data, that is, show more than one or two variables
  4. Completely integrate words, numbers, images, diagrams.
  5. Documentation: thoroughly describe the evidence.
  6. Content counts most of all.

Much of this is true, but uncontentious, and for the most part not very helpful in actually creating a graphic. Many graphics conform with the above principles and still fail to elucidate the meaning.  But the third principle “show multivariate data” is so broad as to be meaningless. Show multivariate data in every graphic? Why? From Tufte’s example of baseball team results, the use of multivariate data often tips the graphic into incomprehensibility, and it would be disastrous if young graphic designers attempted to put such a principle into practice. Intelligibility is all – and that principle, more fundamental than any of the six provided by Tufte, isn’t mentioned. Give me clarity before multivariate, comprehensible graphics before beautiful evidence any day.