A great colour scheme can make your data visualisations more intuitive and more accessible to the viewer. It can set the right mood and draw people’s attention to the most important parts of your data story. A bad colour choice… well, we’ve all been there.
There isn’t really any “magic bullet” when it comes to choosing the right colour scheme for your data visualisation, but there are a couple of general rules of thumb that can be leveraged to elevate your data stories. Here’s what we’ve learned while building our own data visualisations.
Sequential, diverging or qualitative data: what are you visualising?
The first and main challenge when choosing colours for your data visualisations is understanding the data you’re working with and how colours can either enhance or diminish its impact. You got that right, bad colour choice can make your data visualisation difficult to understand or even warp the message you’re trying to get across (no pressure).
The three main categories that we should distinguish here are sequential data, diverging data and qualitative data. Now let’s look at how they differ.
Sequential colour schemes are logically organised from high to low using a gradient effect and need to be represented by sequential lightness steps to form a clear visual message. You’re typically aiming to show progression rather than contrast with low data values represented by light colours and high values represented by dark colours. A gradient-based colour scheme is the best choice to visualise progression.
Diverging colour schemes come in very handy when you want to highlight deviations above and below the median data range. A typical diverging colour scheme is a combination of two different sequential colour schemes based on two different hues with a shared colour at the critical midpoint. Each extreme is represented by the dark colours of a different hue, allowing for a stark visual contrast.
Qualitative colour schemes are typically used to represent nominal differences or differences in kind. By using light and dark hues, you can easily create a lot of contrast and emphasize the most important data points. Keep in mind that the lightness of the hues used for qualitative categories should be similar but not equal.
QUICK tip: To visualise qualitative data, use equally bright hues with contrasting colours. This will ensure you don’t skew the interpretation of your data by drawing attention to one data point over others. To visualise sequential quantitative data, use a gradient-based colour scheme that’s made up of different shades and tints of one hue. This will help you to highlight the progression from high to low.
Where does colour psychology come in here?
Now that we’ve got that out of the way, let’s talk about choosing the right hues. Hues are the unique colours in their purest form, without any shading or tinting -- that’s your blue, yellow and red. Unique hues are critical in data visualisations when you need to create contrast and suggest to the viewer that your data points are comparative, not correlated.
Leveraging the right colours in data visualisation allows us to use the psychological attractions to enhance the visual and complement the story. That’s because the human brain is wired to use colours to discern the world. Because of their psychological effect on our emotions and perceptions, colours are also a critical element of subliminal messaging.
Here’s how our brain perceives the three primary colours:
Red is associated with passion and energy. It can evoke strong emotions and create a sense of urgency. It is also known to stimulate the appetite and act as a physical stimulation, accelerating the heart rate, nerve impulses and blood pressure. In data visualisations, red is the go-to colour for creating emphasis and drawing attention to a particular data point.
Yellow taps into the part of the brain that deals with logic and enthusiasm, that’s why it’s often used to stimulate impulse purchases. It is considered to be a cheerful, happy colour, but too much of colour yellow can cause anxiety.
Blue instils a sense of calm, reliability and security and is associated with maturity. It is also often used to stimulate productivity.
The cultural aspect of colour psychology should also be taken into account if you’re communicating to a global audience. Colour meanings can differ drastically depending on the reader’s cultural background. For example, red symbolises good fortune and prosperity in China but evokes feelings of danger and caution in the Middle East. In Latin America, red is the colour of religion when used with white.
Takeaway. Colour selection in data visualisation is not merely an aesthetic choice, it plays a crucial role in conveying quantitative information. A wrong colour scheme can distort relationships between data values, whereas properly selected colours convey the underlying data accurately and help to tell data stories more effectively. When selecting a colour scheme for your data visualisation, consider the psychological effects of colours to avoid evoking unwanted emotional and psychological responses from your audience.
The simple rules for choosing the right colours
- Have a wide range of hue and brightness.
The whole point of searching for appropriate colour palettes is to make the data more accessible and easier to distinguish. And that can’t be achieved without enough variance in brightness and colours. While brightness in monochromatic palettes can make a huge impact on how accessible the data becomes, it may not always be enough. Consider also adding different hues to make it easier for viewers to distinguish data.
Monochromatic colour palette
- Keep it natural.
Let’s not forget that our brain is preprogrammed to favour natural colour palettes that we see in nature all the time. Here is a great example from Graphiq; a colour progression that transitions from light purple to dark yellow isn’t at all as pleasing to the eye as a transition from light yellow to dark purple. That’s because we’re used to seeing vibrant yellow morphing into deep purple in beautiful sunsets, whereas the former transition feels disharmonious and unnatural. If you were to use such colour palette in a data visualisation, you would lose the reflexive intuitiveness that comes with natural colours and makes it harder for viewers to comprehend the data.
Image source: Graphiq
- Consider the colour-blind people.
Approximately 1 in 12 men and 1 in 200 women are affected by colour blindness. While most colour-blind people can see things as clearly as everyone else, they’re unable to fully see the red, green or blue light. The most common form of colour blindness is known as red/green colour blindness, which means the sufferers mix up all colours that have some red or green as part of the whole colour. To work around this problem in your data visualisations, avoid using red and green together and if you must use them, leverage light versus dark. Ideally, you should try to go for a colour-blind-friendly palette, such as blue & orange or blue & red.
Normal vision vs. Red/green colour blindness
Our colour palettes in action