Effective data visualization takes more than beautiful design
June 23, 2016
By Duncan Angove, Infor president
The data economy has created an insatiable demand for advanced analytics and easily consumable content. But too many visual executions are distracting, or worse, pointless. Harvard Business Review’s Scott Berinato offers a thorough look at the state of data visualizations, breaking down the typologies that distinguish one illustration from another, and revealing why some fail—even when they’re visually compelling.
Even a vaunted publication such as Mary Meeker’s annual Internet Trends report can fall victim to some bad data viz habits. While several sites rushed to create bite-sized articles out of the most essential slides in her recent presentation, tech curmudgeon Josh Bernoff preferred to highlight the ones he found to be the less successful.
The problem is often a matter of putting the cart before the horse—racing to translate data through design and a visual vocabulary before stopping to ask what’s the goal? Berinato proposes a simpler starting point:
“You should begin by asking two questions:
- Is the information conceptual or data-driven?
- Am I declaring something or exploring something?”
The goals of your visualization project will depend on how you answer. Conceptual information means you have a big idea that requires simplification. Making a declaration about the concept will often call for metaphorical illustration; exploring a concept is more a matter of problem solving, wherein working with visual elements helps to generate an idea and shed light on the inner workings of an otherwise complex system. A data-driven project may demand a stronger grasp of basic chart fundamentals, in order to efficiently tell the story behind a set of statistics.
Berinato goes on to share a set of best practices for approaching each type of data visualization, clear guidelines that all designers and data geeks like myself will find helpful. But his emphasis on the importance of outcomes reminds us that the work doesn’t even have to be beautiful.
For instance, Washington, D.C. just published the most detailed lead pipe map ever, and while not particularly attractive, it very effectively points out which urban areas are most at risk of water contamination.
Does that make it a successful data viz? According to Berinato, emphatically yes, because visualization itself is merely a process:
“What we actually do when we make a good chart is get at some truth and move people to feel it—to see what couldn’t be seen before. To change minds. To cause action.”
I continue to be intrigued by innovative data viz techniques, especially drill-down features and other elements that help provide rich context—a tenet of Infor’s approach to analytics. And I’ll take a smart, useful visualization over a pretty but pointless one every time.