When designing a map or a visualisation, sooner or later there is the point where you have to choose a range of colours (except in very specific circumstances which may require you to produce a black-and-white or greyscale visualisation). What is there to consider in such a situation?
Appropriate use of colours
According to Bertin‘s (1918–2010) seminal work, Semiologie Graphique, colour (defined as hue with constant value) as a visual variable is both selective and associative. These mean, respectively, that an object with slightly differing hue can be selected with ease out of a group of objects and that objects with identical colour but differing values for other visual variables (e.g., in the case of shape as the other variable: a red circle, a red square and a red triangle) can easily be grouped mentally. Colour (again: defined as hue with constant value) is, however, not well suited to characterise orderings or quantities. (Note, importantly, that the situation is different for colour value (light or dark), rather than hue. Thus, we see many choropleth maps depicting ordinal values as different colour values, e.g. from bright red to dark red.)
Use of appropriate colours
There are web-based tools that can help with designing colour schemes. I would like to introduce you to three of them:
Colorbrewer by Cynthia Brewer, Mark Harrower and Pennsylvania State University is a tool especially suited for choosing colours for cartographic tasks. It lets you choose the nature of your data from sequential (e.g., temperature), diverging (e.g., political parties) and qualitative (e.g., administrative divisions). Additionally, you can require the choice of colour schemes to be colourblind safe, print friendly and photocopy-able. Colour schemes can be exported manually in different colour spaces, as Adobe Swatch Exchange files and in some more ways.
Comments: The tool is very lean and straightforward to use. It offers a good, but limited, selection of colour-ranges which cannot be adjusted by the user.
The makers of Photoshop, Illustrator et al. offer a versatile tool to view and create colour schemes. Kuler is more complex than Colorbrewer and is not specifically geared towards cartographers or GIS specialists. The tool is social in so far as it lets you view, and choose from, an enormous collection of colour schemes as defined by other users. With Kuler Pulse you can even see which colours are especially popular with the community at the moment. If you rather want to create your very own colour scheme, you can do so using an image as template (Kuler extracts the five main colours) or from a single colour which you choose using a colour wheel. Either way, the number of possibilities to adjust and tune your colour scheme seems endless.
Comments: Kuler is a powerful tool. It is not specifically geared towards cartographic or information visualisation users and gives the user every freedom to choose an unsuitable colour scheme. It is best used when you are not a complete beginner anymore.
Color Scheme Designer (depicted above, found via geoDemesne) resembles Adobe Kuler clearly more than Colorbrewer. Again, the main tool is a colour wheel and a set of configurations such as mono, complement, triad, etc. At first sight, the colour wheel seemed a bit fancy and over-designed to me, but everything seems to make sense: the “steps” in the colour wheel are ranges of distinct colours, which the makers of Color Scheme Designer apparently wanted to promote as base colours. The colour scheme can be exported in HTML+CSS, XML, Text, Photoshop and GIMP formats.
Comments: Color Scheme Designer is a valid competitor for Adobe’s Kuler. An interesting addition (which Kuler lacks) are the colourblind preview modes which also indicate the prevalence of each condition in both genders. A nice touch is the native support of GIMP as a Photoshop alternative.
Making safe choices
When choosing a colour scheme for your map or visualisation, don’t miss the opportunity to make sure your choice is safe for colourblind people. There are various types of colourblindness, the most prevalent of which – Deuteranomaly – occurs in up to 6% of men (and only 0.4% of women). That is 1 in 16 persons and ideally they should be able to understand your maps and visualisations, too, if they sit through your presentation!
Colorbrewer and Color Scheme Designer let you check already ‘out of the box’, if colourblind people can appreciate your colour scheme. What if you would like to use Kuler instead or if you want to test a given colour scheme outside of these tools? You can find a good set of colourblindness websites and tools here (geared a bit more towards (web) designers) and here (for cartographers/GIS specialists).
You can read more about the origins of colourblindness and get more tips for making your visualisations suitable/safe for colourblind people, for example by redundant coding, here.