Colours play an important part in the graph design. Chosen well, they can steer you towards greater understanding by drawing your eye to the most significant parts. Extra colours can act as an additional variable, which makes your diagram more informative. On the other hand, used for the sake of making your diagrams look pretty, they can create clutter, distraction and confusion.
When designing visualisations, we need to bear in mind that the human eye can only differentiate between a limited number of different colours on an individual chart.
First, choose a suitable colour scheme. Think about how your diagrams will be seen by your audience. Will they be viewed on a computer screen, as printed copies or will they be photocopied? Will the colours chosen by you show their distinctions when printed using a black and white printer or when photocopied?
Consider vision deficiencies
If you need to use several colours, consider the fact that red-green colour blindness affects approximately 8% of men and 1% of women, although its severity varies. Red and green combination can be perceived by the majority of colour-blind people as light brown without much distinction, so it’s best to avoid it.
There are three main types of colour vision deficiencies:
- Deuteranopia – colour blindness resulting from insensitivity to green light, causing confusion of greens, reds and yellows
- Protanopia – a hereditary condition and the commonest form of colour blindness, resulting from insensitivity to red light, causing confusion of greens, reds and yellows
- Tritanopia – a rare form of colour blindness resulting from insensitivity to blue light, causing confusion of greens and blues
The diagram below is an approximated simulation of how red and green can be seen by people affected by the first two conditions which are most common. More information and advice on colour vision deficiency can be found on the Royal National Institute of Blind People website.
Familiarity of red, amber, green
Bearing that in mind, it’s interesting to note a widespread use of traffic lights in dashboards as a form of visual alerts.
A popularity of this type of display can probably be explained by the familiarity of road traffic lights and their well understood convention of red meaning danger and green depicting that all is well. What is also very familiar is the actual position of the colours – red at the top and green at the bottom.
This positioning helps to decode the meaning of the information rather than the colour itself. Perhaps a use of a single colour with varying intensity from light to dark or pale to bright, would do a better job for someone who can’t distinguish between red and green.
Don't overdo it
Think very carefully about introducing any extra colours in addition to the base one. If they don’t increase effectiveness of your diagrams, then it may be best not to add them.
Choosing the number of colours representing data categories is an important part of design. Increasing that number will result in a more information rich display by decreasing the amount of data generalisation. However, too many data categories may overwhelm the reader with information and distract them from seeing general trends in the distribution. In addition, a large number of colours may compromise legibility when colours become increasingly difficult to tell apart.
From our experience we tend to find that you should use up to a maximum of five to seven colours to make them easier to distinguish.
You should choose your colours with care, using pale colours for areas (such as bars) and more intense colours for dots, lines, and anything you want to pick out or emphasise. Very bright and clashing colours are difficult on the eyes.