Whenever we teach data visualization at some point a discussion arises where students ask should I put my time into making my visualization look nice, fancy, sophisticated, … or should I focus on something simple that does the job? I always argue that something simple that does the job is the way to go. I often quote Jason Fried:

“Be clear first and clever second. If you have to throw one of those out, throw out clever.”

Or paraphrase him in my own less elegant way:

“If your solution is not based on bar charts and line charts, think carefully!”

Almost inevitably, however, someone with a design background will challenge this position. And they are right to. Anyone who has struggled to grab attention for their visualization work in a magazine, newspaper, or website knows that something plain, no matter how effective, may not manage to grab the fickle reader’s attention long enough to put this effectiveness to use. And so we land back on an old argument of form versus function.

One of the most high profile instances of this argument (in data visualization circles at least) was between Stephen Few and David McCandless around McCandless’ Colours in Culture visualization. Few made a strong argument that function had been sacrificed at the altar of form and demonstrated very clearly that a simple grid visualization as opposed to McCandless’ more visually arresting circle actually did a much better job (see images below). In spite of this being an old argument it continues to raise its head even now!

Source: David McCandless http://www.informationisbeautiful.net/visualizations/colours-in-cultures/

Source: Stephen Few http://www.perceptualedge.com/articles/visual_business_intelligence/our_fascination_with_all_things_circular.pdf

We think that one of the best writers on this topic is Alberto Cairo, who marries deep expertise in data visualization with the experience of working in the publishing industry for years. I think that Cairo’s books, The Functional Art and The Truthful Art, are the best sources for discussion of good data visitation that remains visually arresting. Having read these books I think I will now respond to the design students in my classes with a quote from Alberto:

“A data visualization should only be beautiful when beauty can promote understanding in some way without undermining it in another.”