Choosing any color scheme -- whether for graphics, websites, brands, etc. -- is a challenge in and of itself. That choice of colors sets the mood for anything and everything you create.
When it comes to data visualization, color is especially important. The color scheme sets the tone of the imagery and each color serves to represent a unique piece of information.
The colors you use in your data visualizations represent more than just one idea. The color scheme you choose has the power to display the type of data you're showing, its relationship, the differences between categories, and more.
This post will take you through the process of choosing the perfect color combination for your next data visualization -- from understanding your data to finding the right color tool.
The first step when choosing a color scheme for your data visualization is understanding the data that you’re working with. There are three main categories that matter when choosing color schemes for data: sequential, diverging, and qualitative color schemes.
Sequential color schemes are those schemes that are used to organize quantitative data from high to low using a gradient effect. With quantitative data, you typically want to show a progression rather than a contrast. Using a gradient-based color scheme allows you to show this progression without causing any confusion.
Diverging color schemes allow you to highlight the middle range/extremes of quantitative data by using two contrasting hues on the extremes and a lighter tinted mixture to highlight the middle range.
Qualitative color schemes are used to highlight -- you guessed it -- qualitative categories. With qualitative data, you typically want to create a lot of contrast, which means using different hues to represent each of your data points.
Note: The images above are from Color Brewer 2.0 -- a data visualization color tool designed for working with data mapping. Check it out for your next data map visualization or for grabbing pre-made color schemes to use based on the sequential, diverging, and qualitative models.
Now that you’ve determined which kind of display you want to use, it’s time to determine the number of hues you need to use.
Your hues are the unique colors (like red or blue) in their purest form (without any tinting or shading). Using unique hues is what creates contrast. In data visualization, creating contrast is highly important because it tells the viewer that the contrasting colors are comparative data points. Contrasting colors suggest that the data points are categorical, not correlated, showing you the difference between them rather than the relationship of progression.
Keep in mind that it's possible to use both a sequential and qualitative color scheme in the same visualization. And if this is the case, you’ll need to build a scheme that uses both gradients and unique hues.
One very important tip for creating and finding color schemes for you data visualizations concerns understanding and utilizing the brilliance of colors for a purpose.
In the two pie charts below, notice the brightness of the colors used. On the left pie chart, you can see that there are four main hues used and four tints of each hue. This might signify a relationship between the hue and the tints, or it may just be used to draw attention to some sections of the data over the others.
On the right pie chart, all of the eight hues used have the same brightness. None have more white or black added them to create a shade or a tint, which ultimately creates a balanced, contrasting aesthetic.
Shading and brightness is incredibly important to consider when creating data visualizations because it can be easy to skew the interpretation of your data by drawing attention to some data points over others.
For qualitative data, unless you’re trying to point out one specific data point’s significance, try to use equally bright hues with contrasting colors to display your data.