In a previous post I referred to the fact that visualization is based on a vocabulary of marks and symbols and a grammar that informs how we transform them to encode data and how we layer them. This gives visualization its flexibility and power. However, we need to acknowledge that sometimes it's not clear how to read a type of graphic, particularly if we haven't seen it before. If you use visualization in innovative ways and don't explain how to read it, your graphic may be ambiguous or hard to understand. Annotations matter.
Take these mesmerizing full-page visualizations published by The New York Times today:
Here are some detail photos:
It took me a while to realize that the feature that is encoding time isn't line length, but depth: lines here are roots of a tree; therefore, what encodes time is the distance from an imaginary horizontal ground line —where Google and Facebook are on each graphic— to the tip of each root.
What about the horizontal sorting of the lines? The online version of the project includes a critical feature that is absent on print: a X-axis scale that reveals that horizontal position corresponds to the number of acquisitions per year:
I've been reading and making visualizations for decades; if it required so much effort on my part to grasp what these graphics show, try to imagine the frustration that people with less experience may feel.
Important aside: there's another myth I've been fighting against for years: “readers must be able to understand any visualization in a few seconds.” This isn't true. It's also self-defeating, as it may lead us to design simplistic graphics, and to employ just common forms such as bar or line graphs. Visualizations ought to be complex sometimes —just because the stories they tell are also complex— and, like good writing, they demand your time and attention to yield insights. Also, it's appropriate to experiment with novel graphic forms if only to expand our (and our readers') visual vocabulary.
However, we could make people's lives a bit easier by telling them what they're seeing (show AND tell!) This is just a suggestion: what if we added a small how-to-read-this-graphic sidebar somewhere?