Estimated reading time: 2 minutes

Summary

Aesthetic choices matter when you're choosing colour, particularly when you're communicating with an audience that lacks subject or discipline-specific knowledge
Even if you use a pre-existing palette, make sure the colours you've chosen fit your data, contain enough contrast, and that you have enough colours to choose from so that you can best present your work

After taking into account established practices, discipline-specific preferences, physical conditions and cultural biases, aesthetic choices still need to be made. This is particularly relevant in cases where the purpose of the visualisation is to communicate to an audience that lacks adequate subject knowledge, discipline knowledge or insight into the scientific basis of the visualisation. This might include colleagues outside the research group, grant applications, fellow students in the same seminar, scientific publications, or popular science communications.

It is the nature of aesthetics to be sensitive to trends. Opinions on what is most appealing come and go. However, the colour selection process is relatively permanent. In the recommended resources section below you'll find a collection of articles that describe the colour selection process from a perspective relevant to science communication.

Pre-existing colour palettes

In some cases, there may be a predetermined palette to adhere to as part of a graphic profile. This may involve an organisation's brand colours, a publication's colour theme or the limitations of a presentation template.

Karolinska Institutet, for instance, has a graphic profile that KI staff are meant to comply with when they produce visual material for communicative purposes. It has five primary colours in addition to black and white, and there are specific colour guidelines concerning infographics and figures (for example bar charts).

Example of colour distribution meant to indicate Karolinska Institutet’s overall colour expression.

In your own visualisation work, always think first about if and how you want to apply the colours provided by the pre-existing palette, and to what extent you intend to use them; it is not certain that pre-existing palettes and colour guidelines are appropriate in relation to the purpose of your visualisation. It can be convenient to use an existing palette and not have to make your own colour choices, but if the palette you use is too narrow, restricted or non-accessible, you run the risk of your visualisation not communicating what needs to be communicated. Consider the possible negative consequences of using colours that do not fit your data, break conventions within your scientific field, or are unaccessible for people with for example colour vision deficiencies.

One recommendation might be to use profile colours for theme-bearing elements, such as the headline or header/footer, but to try to produce a separate and complementary palette of colours that fits your data, is sufficiently generous in terms of colour, and has enough contrast to clearly and accessibly convey the message you are trying to communicate. Then try to stick to that palette when creating any additional visualisations for the same presentation material.

Recommended resources

On choosing colour:

Robert Simmon, Subtleties of Color (2013).
Lisa Charlotte Muth,
- What to consider when choosing colors for data visualization (2018).
- Which color scale to use when visualizing data (2021).
Chrisanthy Rebecca Surya, How to use color in data visualization, Medium (2019).

Useful tools:

Cynthia Brewer and colleagues have developed a tool to help you generate colour scales: ColorBrewer 2.0.
Eric D. Kennedy has developed a similar tool aimed at quickly and easily creating a colour scale: Data Color Picker.

If you want to know more about profile colours and complimentary palettes:

Amy Cesal, How to Create Brand Colors for Data Visualization Style Guidelines, Nightingale (2020).
Cynthia Brewer, Color Use Guidelines for Mapping and Visualization