Don’t just visualize data— communicate it
I enjoyed reading 3 Common Mistakes in Data Visualization by Joanna Ngai. In the piece, she points out that too often, data visualizers make these three mistakes: misusing charts, visual clutter, and busy graphics.
At a glance, the second two seem like they could be similar, but Visual Clutter is more like “too much design” (setting aside the fact that good designers will point out that good design involves as much taking away as it does adding), and Busy Graphics is more like “too much data.”
Both of these hinder many data visualizations from doing what data vis should do: communicate data. To many, “data visualization” refers to hyper-designed, data-dense pieces. A Google image search for “data visualization” yields a quick scan of what I’m describing — big, elaborate pieces with lots of visual appeal.
Few of these make it clear what they’re showing. For many of these, a user would have to take as much time scrutinizing the visualization to make sense of it as they would if they were just looking at the raw data. We know from web usability that we have 10 seconds to capture a user’s attention. Why should data visualization be any different?
As tools to make visual, web-based data features become more common, this has started to shift: I’m seeing more data visualizations that let users explore the data, switch series, and get interpretive blurbs. Choose your own adventure stories, if you will.
However, the conversation around “data visualization” is still more heavily influenced by design — or, perhaps, by “designy-ness” — than by a focused priority on communicating something about the data.
Find the story in the data — and then tell it
Previously, in Public health: write more better, I wrote about the need to make information clear so that our audience can understand it and use it. Data visualization is both a tool we can use to do this, as well as a field of its own that can learn some of the same lessons.
Below is a before-and-after, a screenshot from datavisblog.com, which serves as a quick reminder of why we visualize data. Expecting a reader to sit down and interpret a handful of numbers — even a small, simple data set — is a big ask.
Visualizing these data lets you communicate them more effectively — and it lets the data support the story you’re trying to tell.
The point of data visualization is not just to reformat the data. It’s to use visuals to tell a story that is obscured when the data are raw data. Find the story, and then tell it.
Data vis doesn’t have to be complicated
In fact, it shouldn’t be. Simplification can offer so much. As an example, I often look to the New York Times for guidance, techniques, and examples of how to effectively visualize data for the purpose of communication.
Let’s take a look at what they do to make such clear, powerful charts, graphs, maps, and other data visualizations. They:
- Find the story in the data, using the data to highlight a main message in the story around it
- Reinforce the message with design — like with the red bar
- Interpret or explain the data with a clear title (“A Rise in Attacks”)
- Show restraint with labels: the subtitle labels the units (a big departure from academic standards), axis labels are minimal, and not every value is included
- Simplify and de-ephasize the source and eschew notes
You’ve got the power — so use it
You can create important, clear, powerful data visualizations. don’t need to be a graphic designer, a data scientist, or know anything about “big data.” All you need is MS Excel, a copy of Tufte, Stephanie Evergreen’s checklist, and a willingness to experiment and learn.
Go forth.
