A data visualization should only be beautiful when beauty can promote understanding in some way without undermining it in another. Is beauty sometimes useful? Certainly. Is beauty always useful? Certainly not.
- —Stephen Few, from “Should Data Visualization Be Beautiful?”
The first step to finding the middle ground between radical minimalism and a more playful approach to information graphics and visualization is to remember that a good graphic realizes two basic goals: It presents information, and it allows users to explore that information. In other words, an information graphic is a tool for the designer to communicate with readers, and a tool for readers to analyze what’s being presented to them. It doesn’t matter if you see yourself as an engineer or as an artist: If you create infographics and visualizations, the balance you achieve between these two dimensions will define whether or not your work is good.
It’s Not the Style, It’s the Content
Figure 4.1 contains a graphic I made about the economic performance of the Games&Toys Company (which doesn’t exist, of course). If we apply the visualization wheel (see Chapter 3) to it, we can see that it’s pretty dense, albeit one-dimensional. It gives just one piece of information—how Games&Toys’s after-tax revenue changed between 2006 and 2011—but lets you do little else, other than enjoy the bright colors and irrelevant eye candy surrounding the data line. The graphic presents information, but barely allows exploration.
Figure 4.1. Adding tons of special effects to a graphic will not make it any better if it lacks good information.
Now see Figure 4.2 which shows three scatter-plots I designed based on data taken from The Spirit Level, by Richard Wilkinson and Kate Pickett. The book’s main thesis is that income inequality is related to negative socioeconomic indicators such as pregnancies among teenage women, rates of homicide, mental disorders, and so on. These two plots tend to the other side of the visualization wheel spectrum: They are spartan in appearance, very abstract, and encode a lot of data. In this case, while the charts also present facts, their main goal is to allow readers to visualize relationships and perhaps compare their own country with others.
Figure 4.2. Inequality is correlated with negative socioeconomic indicators.
The differences between Figures 4.1 and 4.2 are deep. The designer who made Figure 4.1—that would be me—assumed that readers don’t care about information if it’s not surrounded by bells and whistles. On the other hand, when I designed Figure 4.2, I assumed that the people reading my work were interested in inequality beforehand.
Do I believe Figure 4.2 is better than 4.1? I do, not because of the decorative nature of the eye candy, but because the special effects take away space that could have been used to highlight other angles of the story; for example, to explain the erratic changes in Games&Toys’s performance. How was the company doing in comparison to its competitors and with the market in general? What caused the surge in revenue: computers and video games, perhaps? Could I have visualized the covariation between Games&Toys’s after-tax revenue and the penetration of digital games at home?
Figure 4.1 is not a so-so graphic because of its style. The reason it’s not good enough is because it wastes too much real estate (and the designer’s time) on things that don’t help readers understand the figures.