Wednesday, January 23, 2013

Andy Kirk's data visualization book

(Disclaimer: I was a technical reviewer of this book. Even if my contribution to it was very limited I helped a little bit with matters of writing and style my name is in it, so take my endorsement with a grain of salt.)

It seems that every consultant in visualization and infographics feels the urge to hold one or more books under her or his arm at some point. Think of Cleveland, Few, Robbins, and now Andy Kirk. Andy, who runs the VisualisingData website, has been very active in the past few years, speaking and conducting workshops in Europe and the US. His first book, Data Visualization: A Successful Design Process,* reads as a detailed distillation of the ideas that he presents in his courses. If you attend one of them, the book may work as a reminder of what you learned.

I won't bore you with the expected but unavoidable stuff: Yes, there's some intro material about what data visualization is and about how it is supposed to work; yes, there are quotes by Schneiderman ("visualization gives you answers to questions you didn't know you had"), and other usual suspects; yes, many graphic forms and the "successful design process" of the subtitle are presented; yes, there's a discussion about software tools, websites, and other resources in the second half of the book. This is a book for beginners, after all, as Andy himself has recognized.

But professionals will stumble upon things to learn or to remember, as well. What I found more interesting in the book myself is the way it addresses some current challenges in visualization. First, the importance of having an editorial focus. Here you have some lines from pages 55 and 56:

"An editorial approach to visualization design requires us to take responsibility to filter out the noise from the signals, identifying the most valuable, most striking, or most relevant dimensions of the subject matter in question (...) You might think the idea of telling stories is only relevant for explanatory pieces. That's not the case. With exploratory designs you still need to demonstrate this editorial focus. The difference is that with these projects you are not so much telling stories rather you are making them accessible and discoverable. You still need to frame the subject matter and define the important dimensions of analysis that will be made available for manipulation and interrogation. You still need that level of care for the audience's interpretive experience."
These are very relevant words indeed (I am a journalist, so I am favorably biased). So are some other parts of the book, such as the section on how to conduct a simple visual analysis of a dataset using graphs, and the nice overview of the annotation layer, a feature that news infographics designers find almost trivial because we use it all the time, but that visualization designers with a scientific background and statisticians sometimes approach with distrust and tend to neglect.

Andy also writes about his now famous
eight hats of data visualization idea, an accurate portrait of the multidisciplinary nature of the field nowadays which, unfortunately, too many managers don't embrace, as they cling to the rotten notion that infographics and visualization are matters that just designers and developers should get acquainted with. Finally, as a funny side note, no other book I know has called the great Jacques Bertin "a clever chap" (page 89.) To my highly americanized ears this sounds much better than its synonym "fellow". It may be because English is not my native language. Go figure.

(*I wrote "book", but I should have said "e-book". The printed version is black and white. If you are not a traditionalist, like me, get the digital version. It looks much better and it is full color.)