Database Reference
In-Depth Information
Figure 1-6. Effectiveness of data encoding
A few points are immediately obvious:
▪ Position is the most effective form of encoding for all data types.
▪ Length, angle, and area decrease in effectiveness from quantitative to ordinal to nominal.
▪ Color hue increases in effectiveness from quantitative to ordinal to nominal.
Keeping this ranking in mind as you select your visualization type will help ensure you are
crafting a message that will be easily decoded and understood by your audience.
If the overall quality of the communication were only affected by the ease of decoding, we
would not need any more principles. In actuality, we also need to consider aesthetics, media
and channel, and the actual impact.
Principle #4: Design for Aesthetics
Let me play devil's advocate: Why consider aesthetics at all? Isn't any attempt to make a
visualization “look better” just chart junk or design fluff? Won't graphic elements that aren't
data just get it the way? Shouldn't the data itself be beautiful enough for readers?
I understand this viewpoint, I really do. I've seen plenty of attempts to beautify data visualiz-
ations that either distract the audience or, worse, distort the data so as to completely mislead
the audience. We all agree that this result must be avoided. One way to avoid it is to banish
all aesthetic elements forevermore. And yet, that's not a world I'd want to live in, because
there is a clear value to elegant design and what Willard Cope Brinton called
“judicious em-
