A few years ago I have used Cindy Brewer's ColorBrewer system for picking the right color scheme for graphics, based on experience from cartography. Recently, it has also been made into a R library. In particular, ColorBrewer distinguishes three types of color schemes: diverging, quantitative and qualitative:
Diverging blurs out the mean (appropriate for visualizing normally distributed real variables, or also correlations with color), quantitative blurs out the zero (appropriate for visualizing exponentially distributed positive variables with color), and qualitative makes it easy to distinguish adjacent values (appropriate for categorical variables).
ColorBrewer provides guidelines with respect to appropriateness of a color scheme for print, monitors, laptops, projectors, photocopying and even color blindness (I've once had someone complain about my color schemes after a lecture - just to find out that he's color blind).
I've been frustrated by people who do not use infinite palettes for visualizing data, and RColorBrewer exacerbates the problem by not being able to create palettes of arbitrary size. But, using simple linear interpolation between colors, one can create very appealing infinite palettes that maintain the approximate perceptual linearity (meaning that a change in our perception of color strength is proportional to the change in value across the scale) of ColorBrewer's palettes:
While some people might argue that 11 bins are enough, I'd respond to this by saying that binning is an act of pure and inexcusable laziness in the case when you can easily visualize a continuum.
Just a few days ago, I've come across another tool: Juice Analytics Chart Chooser. It lists a number of charts: both as pictures, and also as PowerPoint and Excel templates:
Each chart may have some of the following features:
- Trend involves a variable indicating time
- Composition involves a set of variables that add up to 1
- Distribution exhibits an occurrence count for different values of some variable
- Comparison pairs up two or more variables
- Relationship visualizes a complex relationship between two or more variables
Thereby, you can see how visualizations carry many parallels to models. Picking a visualization is very much alike picking a good model. I have discussed this before.
For the end, here is a chart I use for assigning dimensions to graphical elements. A full circle indicates a good choice, an empty one an acceptable choice, while the absence of the circle means that the element isn't useful for presenting that particular aspect of the data.
Note, nominal corresponds to the qualitative scale above, quantitative to diverging and ordinal approximately to sequential.
Happy Thanksgiving!
Problem in creating "very appealing infinite palettes that maintain the approximate perceptual linearity" is that different media (different monitors, different paper qualities, different inks, etc.) have different characteristics and unless you are very careful with color profiles and calibration you may get undesired changes in perceptual linearity. For example, red may seem more intense than blue, biasing the visual interpretation. When this is combined with tendency of human vision system to find change points in continuous color gradients and thus finding shapes e.g. in colored maps (segmenting along color change points), "the approximate perceptual linearity" is easily lost.
We have used RdBu divergent color scheme for epidemiological maps showing spatial relative risk. Thus middle color has the special meaning of no practically significant difference in relative risk. With continuous color scheme it depends on the media and observer what areas are observed as red or blue and even hint of red or blue is usually interpreted as significant difference by those who have seen the maps. The binned color scheme with moderate number of bins is a more robust choice, allowing better control of which areas are interpreted as having significant difference.
Hi,
I think it would be useful to have a version of divergence - perhaps called distance? that coloured the difference from central (white) the same for positive or negative deviations from middle.
Or it a deliberate choice not to do that?
If so I would be interested to know why...
Dave
@Aki: Some of the palettes can be taken to other types of media. For example, my red-white-blue one isn't appropriate for video projection according to ColorBrewer. The linearity is a bonus, one can easily convey useful information without perfect linearity. As for the white zone, to achieve mapping of insignificant to white, one might have to stretch the white band and make it wider: it's not a problem of infinite palettes, it's a problem of creating the right mapping.
@Dave: Sure, but one can do this my mapping difference into distance and thus creating a positive variable that can be visualized with sequential palette.
The full circles and empty circles are almost indistinguishable, which is sort of ironic in an article titled "assistance in picking colors."
Or maybe it was intentional, and was a clever test, to see if we realize that there is a bad color choice that needs assistance.
@a: have you tried the color blindness test link? :)
Seriously, thanks for feedback. Yes, I agree it can be improved. But I don't think the problem is in the color choice - the problem is really with the fact that the circle outline is thinner when the circle is filled, which creates some confusion regarding the importance.
Since nearly 10% of males and 1% of females are colour blind, there must be plenty more who don't like the colour sceme, but didn't bother mentioning it. And chances are, there is somebody in an office down the corridor, or a student, who you can quickly show a slide to and ask if all the colours can be easily distinguished. If not, all it requires is upping the contrast between two shades, nothing time consuming.