Week 2: Colours in data - less is more | DS23

by Auguste Navickyte

Week 2 at the Data School started with a theoretical introduction to data visualisation. Before DS, I only used my intuition and personal taste for colours when creating dashboards (I love purple). I did not think how to make colours efficient in my work. I have now learnt though that when it comes to colours - there are many details to consider before starting and finishing off your visualisation.

First, be cautious about the number of colours you use. Rule of thumb - the fewer, the better.  It is best to begin with neutral colours (black, grey, white) and introduce colourful colours only when it is absolutely necessary. If you need to add another colour to your viz, think if you could first utilise different shades of the colour that you have already used.

There are a couple reasons for introducing a new colour though. A well chosen colour could evoke emotions that would contribute to your overall work’s quality. For example, red could be associated with threat or danger, therefore, using it in a viz about forest fires would work perfectly for that. Colours also help to create stronger associations. For instant, blue and orange could be used for a dashboard about The Information Lab. Colours also assist with differentiating between common threads - different teams, political parties, etc., which makes it easier to understand a viz more quickly. Do not forget to label the colours though.

You can use colours not just on graphs, but also in titles, subtitles, tooltips, backgrounds, borders, and more - make them colourful so long as they give extra insights and help to understand the viz better. Another rule applies here: if a colour gives no additional insight, do not use it.

Finally - consider colour blindness. There are 3 million colour-blind people in the UK alone. A very quick tip - try to avoid these colour combinations: green-red, green-brown, blue-purple and green-yellow. If this in turn makes you struggle with colour combinations - do have a look at https://coolors.co/. The website helps you to create colour palettes based on your own preferences, which you can then save on Tableau (see instructions here).

Overall, colour does help to highlight key findings better in your dashboard, and therefore becomes useful beyond just its aesthetic purpose - but use it with careful thought and moderation!