Different theories about color have been developed, but there are some shared features:

• The chromatic circle
• Hue
• Luminosity

Chromatic circle: The chromatic circle is a really convenient way to show up the three so-called primary colors: cyan (a kind of blue), magenta (red), and yellow, and how the other colors are related to each other:

From the circle, we can define two relevant concepts: hue and complementary colors.

Hue: In the theory of color, hues are what are commonly known as colors. For instance, within the previous reproduced chromatic circle, every single component has to be considered as a different hue. We can see that for every given hue there are what we call shades, but how do we obtain them? We simply add black or white to the main hue.

For instance, the following self portrait by Picasso is mainly made from shades of blue:

Complementary colors: Complementary colors are pairs of colors that highly contrast themselves if placed next to each other. There are different theories on how to define pairs of complementary colors. The traditional pairs are as follows:

• Red - green
• Yellow - purple
• Blue - orange

This is all interesting, isn't it, but how do we leverage these concepts when dealing with our data visualizations? We are going to look now at two simple but powerful rules that make use of what we have just learned. We are going to use here a population of records, each showing three different attributes:

• A numerical x
• A numerical y
• A category, from A to D

Use complementary hues for data pertaining to different families: As shades of the same hue can show and communicate commonality, complementary colors can communicate differences in groups of data. If we are representing data from different categories of a given attribute, we can leverage complementary colors to highlight their differences:

Use the same hue with a different shade for data pertaining to the same family: When dealing with data related to the same group, for instance, all the time series data related to region within a country, a good way to represent their shared pertinence to the group is to color them with the same hue but different shades. This will immediately and effectively communicate their commonality and easily distinguish them from the other groups.

The following figure shows an example of this concept. You will notice that, even without knowing a lot about the data represented, you will be inclined to assume that A-C and B-D are in some way related: