How people perceive colour is an interesting issue. The Young-Helmholtz Trichromatic theory of vision proposed that we have red/green/blue receptors, which are then combined to show different colours. It is thought that the red-green receptors are close together, and perhaps this is the root of the issue. Needless to say, the issue is complex but interesting.
According to popular wisdom, it is thought that the red-green chromatic channel developed in order to provide an evolutionary advantage for determining ripe fruits against a background of foilage. Tell that to your children, next time they refuse to eat fruit! However, this ‘ripe fruit’ theory has been difficult to observe in field studies. One group of researchers conducted field studies in black-handed spider monkeys, and found that luminance contrast was just as important in distinguishing fruits. If you’re interested to read more, here is an interesting study that illustrates the complexities of perception, which involves the field study of primates. On the other hand, a separate study showed that trichromatic primates found it easier to determine and select ripe fruits, and you can find more information here.
How does this impact data visualisation? It is possible to produce visualisations that make the most of luminosity in order to encode values, along with the size of the data point, in order to convey the message of the data visualisation. Another issue is that determining colour and luminosity can be a subjective issue, and point size may help to provide additional cues. I envisage it as if it is the detection of fruit in viticulture. Therefore, one winemaker might ascertain that a grape’s optimal point of ripeness is at one point, and another viticulturist might determine that the ripeness point is at another point in time. Similarly, it isn’t always easy to ask experimental subjects to ascertain the amount of ‘greenness’, ‘redness’, or ‘blueness’ of a point. There has been some work in computer vision, aimed at distinguishing the RGB in fruit, which is interesting to read.
It is suggested that about 12% of males are colour-blind, which means that they are restricted from using the red-green channel. If you are interested in reading more about the experience of a colour-blind person, please do read this entertaining blog by Geoffrey Hope-Terry.
To summarise, data visualisations can therefore augment understanding by assisting the perceptual processes involved in luminance and the blue-yellow colours. It is also possible to use the size of the data point to convey the message of the data. In other words, data visualisations should aim not to exclude members of the audience by including lots of red and green together.
3 thoughts on “Color-blindness – why does it happen, and how can data visualisations help?”
It’s good that you bring attention to the fact that not all see colors that way we do. However, avoiding red and green alone does not solve the problem. According to the Lighthouse International, “It is important to appreciate that it is the contrast of colors one against another that makes them more or less discernible rather than the individual colors themselves.” They go on to stress the importance of varying lightness and saturation. Most people with color vision deficiencies can distinguish a light red from a dark green or a dark red from a light green. There are also other color combinations that cause trouble for some.
Further to the above, I think this is an important and often misunderstood point. As someone who has been diagnosed as red/green colour blind it is not the case that I can't see the two colours clearly. Whilst this may be a problem for some it isn't the experience of most of us.
Green and Red do appear as clear and distinct colours, the problem comes when I try to distinguish between them in close proximity, for example small dots of these two colours tend to look the same. So in the classic tests where a number jumps out to most people, for me it is camouflaged into the background.
In my case at least, please don't imagine that a red car doesn't appear “red”!
When it comes to data visualisation, I have never found myself unable to read a chart. However it is true that red and green bars or data points don't leap out in the same way as others colours might.
I wanted to thank you both for your insights into colour blindness, and for taking the time to comment on my blog. Although this was just a short blog, it became clear to me that this issue is extremely complex.
The interesting thing for me, in terms of data visualisation, is the element of subjectivity and individualism.
Data visualisation is, for me, analogous to John Stuart Mill's Utilitarian Principle of maximising the most amount of happiness to the maximum number of people. In other words, we are trying to make the visualisation as clear as possible, to as many people as possible.
I'm not sure how light red and dark green would look together in a chart, however. It may come down to balancing aesthetic considerations too.