I get it: a particular data chart type is absolutely right -- and another type is wrong -- for any given application. I too love to debate the correct level of precision for any chart type, and whether a series reflects a stock or flow variable.
But there is another legitimate area of visualization which isn't about the data itself. The good folks at Visual-Literacy.org (http://www.visual-literacy.org/periodic_table/periodic_table.html) provide several names for sub-types, but for my purpose today I lump them all together as "Concept Visualization." If you are a data geek, you should label it "Here Be Dragons" because it is likely to be off your map.
Concept visualization is not susceptible to the math and logic arguments that define the correct selection of data charts, the correct use of series colors or overlaps, and the correct precision level for reporting. Concept visualization is a soft science. But I bristle at the thought that any communication method is devoid of best practices or motivating concepts.
These four concepts seem to me to be solid drivers of best practices when you leave the comfort zone of data visualization:
Ambiguity is Absolutely Sometimes Bad
We don't want to confuse our audience with a non-data chart that looks suspiciously like it contains data. Concept visualization is all about that which isn't data. Some chart types are popular because they seem cool and edgy, but violate this principal. When you look at them, you naturally want to analyze them numerically. Witness this Radar Chart (thanks to Visual-Literacy.org).
It looks like data, doesn't it? And it is in no way untrue. Can't you imagine a manager wanting to calculate the area inside the chart values?
I consider radar charts guilty of data ambiguity. I think there are better ways to demonstrate a set of scores or values when you don't have actual data (See my Parameter Ruler, at the end of this blog).
Concept visualization as Cartographic Metaphor
Concepts are often about relationships, flows, and decisions. We easily consume these matters as maps: geography, roads and intersections. The Data Flow diagram at the left is a classic example.
Instead of a flat chart of reality, it looks a lot like the island treasure map from a children's book. It is visceral, direct and memorable because of this. Its failure to be literal is not a weakness but a strength.
The danger is again that we confuse our audience with designs and arrangements that look literal but aren't, or that we fall into our habit of making everything technical and data driven.
Slick and Non-stick are Opposites
It is all too easy these days to muck up a presentation with silly, distracting frills. I don't encourage that. But it is perfectly easy to make your charts slick and professional: 3D formats, bevels, lighting, gradients, and shadows. Judiciously and consistently used, these add the professionalism that helps a concept stick where it is intended: in the minds of your audience.
Stranger than Fiction
Remember that concept visualization can merely be refreshing, clear and memorable. So the fact that they aren't literal, and are even a little screwy, can be the key to their value. The internet is full of metaphor charts that are pungent and memorable specifically because they are not quite true. For instance, an iceberg chart illustrates nicely the relative risk associated with identified risk compared to unknown risks (below the water line), even though few of our projects take place in the North Atlantic during winter.
The image above is a parameter chart I built in Excel. It is automated, and shows scores on five dimensions by sliding strips of paper back and forth inside a ruler. It isn't literally true in any way. But by breaking people out of their every day modes we make them accept information, concepts and relationships more openly, and the very oddness of the metaphor, map or image can be the hook by which they recall your important information later on.