Once upon a time, in a far away land called New Haven, CT, there lived a man who would revolutionize the way analysts of all breeds communicate data. A colleague from my past introduced me to Edward Tufte (hereafter: ET, his self-proclaimed nickname) several years ago. ET is on a mission to make the world a better place, one chart at a time. Specifically: he has developed a philosophy toward design that seeks to maximize the flow of complicated sets of information from the medium (design space) to the end user (you and me).
When I saw that ET would be teaching a workshop here in NYC, I jumped on it. I was (mistakenly) under the impression that ET was singularly a master of jamming as much information into a single chart as intelligibly possible. After all–who wouldn’t want to pick up a few good templates to add to the reporting arsenal? While we certainly learned tactical approaches to designing data outputs, ET conveyed a much richer appreciation and respect for the process of translating data into visual displays. The crux of his message is that data should drive design–and not the other way around.
The example above (Napoleon’s March – 1812, which was presented by ET) demonstrates how complex information can be amassed for maximum impact. The image overlays 1) a time series of Napoleon’s army size, 2) geographical elements conveyed via a map, and 3) change in temperature. Though it may appear overly complex, the image is pretty straightforward: it provides evidence as to why Napoleon lost so many men. As the road became more treacherous (crossing rivers) and colder, more men died. Spend a few minutes absorbing the data and I think you’ll note its elegance.
In a world where we feel compelled to visualize everything, it’s no surprise to see cartoon-like charts that are 100% style with no substance. Just because you can visualize something doesn’t mean you should, especially if you can convey your point more directly via some other solution. A pretty chart with slick graphics and lots of modules that lacks any relational depth between numbers or dimensions is NOT an insight tool–it’s just a pretty chart. And design can’t work magic where there is no meat to be found. As ET puts it:
“If your numbers are boring – get better numbers.”
(Stepping down from soap box now).
It must be said that the ET experience itself was quite remarkable. There must have been ~1,000 professionals from a host of industries–private and public sector, commercial and academic, marketing and otherwise–jammed into a huge auditorium and ready to learn from the master. We were treated to a day-long lecture that was rife with design theory, hands-on examples, multimedia, and anecdotes that actually elevated the points being made. While it would be impossible to do the day justice with a single blog post, here are the broad theories that anyone tasked with creating Visual Displays of Quantitative Information must consider:
- Be prepared to do “whatever it takes.”
The right question to ask: “How can we best understand the data?”
The wrong question to ask: “How can we use visualization to understand data?”
Zero out the interface and make it about the content. Find the most intuitive point of connection between your communication platform and those who will consume the information.
Always put data in context. You can achieve this by including comparisons, causality, adjacency, multivariate dimensions, and always–lots of documentation.
- Finally (for this post): the design logic of the display must equal the intellectual logic of the analysis. In other words, do right by your information by designing a relevant presentation format–and don’t be afraid of “data density” if your analysis begs for inclusion of multiple important factors.
If you are interested in learning more about the world of ET, check out his website here.Share