Martin Krzywinski on How Graphics Can Save Lives
Newton encapsulated how science works: You try things out. You communicate what you find. And you repeat.
Now, unfortunately, when scientists communicate, wildly confusing and unintelligible presentations and graphics can be created. I liken this to a disease, a communication disease like polio, which ideally should be eradicated.
Design loves science. And science also loves design. The problem is, science doesn’t always know it loves design. Scientists are very poor at communicating visually, but they’re also very poor at knowing this. Now we might think, “Well, why? They’re smart. They’re trained. It’s not rocket science after all” (not to disparage design). But it’s exactly because it’s not rocket science that they can’t get it. If it were rocket science, we’d be on the moon.
Here’s an example. In a study published in the BMJ medical journal, doctors were asked to look at four kinds of visuals, and they were asked to rate the visuals for two things: Were they correct—were they accurate in interpreting the information—and did they have a good time? Because you should always be asking that second question.
The scientists chose an icon graph as the most accurate representation. Nobody liked it. They were right, but they didn’t have a good time being right. They had other forms of communication in their head that they thought would deliver the information to them better. Moreover, they were contemptuous of the accurate-but-unenjoyable graphic. It was like when you bypass the broccoli at the grocery: I know this is good for me, but it’s not for me.
We need to be aware that we’re in love with certain things we do not need to be in love with. There are visualizations out there that look good but are not really an effective means of communicating information—they rely on little more than optical illusions. So when we’re looking at something like, say, a network visualization (more commonly called a “hairball,” a graphic that’s popular right now), we should ask ourselves, “Am I being fooled into thinking I know something? Am I being fooled into thinking this is telling me something?”
So what do we need to do? I think we have to encourage and reward effective visual communication. When you see a scientist trying, you should reward them. And when they do a good job, you should really reward them. Second, we need to train scientists to do a better job.
We also need to show how design can help embrace complexity. Complex things are hard to understand. We have to make that complexity accessible.
Not that long ago, we thought big data was around the corner. And we thought, “Well, we’re going to need computers to help us number-crunch this stuff, because the input is too complex.” But what happens now is that, much of the time, the output is equally complex. Computers can help us process the data; design can help us express it.