In all of our workshops (including our virtual ones), we include time for free-form questions and answers. One topic that often comes up is how to use color effectively. We spend a lot of time talking about sparing use of color, making sure that we're using it to focus our audience's attention where we want them to pay it. But that usually leads to a related series of questions:

How do we pick the “right” color for the specific visualization that we are creating?
What is your advice on creating a visually pleasing palette of colors?
What if I need multiple distinct colors? How do I make sure that one stands out more than the others?
Do I choose a different set of colors based on the relationships among the elements in my chart?
How can we use color effectively, when we are required to use the color that goes along with our corporate brand?

We know how to use color effectively and sparingly, but sometimes the challenge is in picking effective and appealing colors for the message we're trying to get across.

Good news, everyone: we can use the color wheel, and a few simple guidelines, to help us select an effective set of colors for just about any visual we create. In this post, I’ll talk about:

What terms like hue, saturation, lightness, and temperature mean
How a color wheel works (and where to find it in your Office applications)
Techniques for choosing a key color for your palette
What “color harmonies” are, and which ones work well for comparing two things, three things, or four things
How palettes for categorical data differ from those for quantitative data
Some online resources for easily creating your own harmony-based color palettes

A quick explanation of the components of color and the color wheel

What we think of as “color” is made up of a few distinct components: hue, saturation, and lightness. Visualizing different colors on a wheel helps us compartmentalize those components in our mind, and will make it easier for us to pick an effective palette for ourselves.

Hue is best thought of as “what color of the rainbow is this?” On a color wheel, we plot hue around the circumference of the wheel.

Saturation is how bold (very saturated) or how pale and washed out (very unsaturated) a color is. On a scale of 0-100% saturated, 0% would be gray, and 100% would be a pure bold version of that color. On a color wheel, we show saturation on the radius of the wheel. Closer to the center is less saturated, and closer to the edge is more saturated.

Lightness has a few different definitions, but the gist of it is: the more white you add to each color (also called “tinting”) in a color wheel, the higher the lightness goes, up to 100% (which would simply be white). If you added black instead (also called “shading”), the lightness would decrease towards 0%, which would just be all black. Most interactive color wheels have a slider below the wheel itself to adjust the lightness. Our sample color wheel has a lightness of 50%; you can see what color wheels with different lightnesses look like in the chart below.

Where can I find the color wheel in my applications?

When you open up Excel or PowerPoint, unfortunately, your color choices aren't presented to you in a color wheel, they're presented to you in a pre-selected palette. Those strips of colors along the top represent the hues of that palette, and then the columns of related colors below it represent shades and tints of that color (modifications of the lightness).

Fortunately, you can access the color wheel by clicking on the “More Colors…” button in the pop-up box that appears when you click on the paint bucket to fill a shape, or the pen to outline a shape.

You’ll see a color wheel that uses a smooth gradient: the hue changes as you go around the circle, the saturation decreases as you go towards the center, and the slider below the wheel changes the lightness. If you can visualize color harmonies as though they are on a clock face, you should be able to pick them out using the MS color wheel for any key color you like.

Do colors have temperatures?

Temperature is not an inherent property of colors, but rather a characterization of how we perceive them. We think of some colors as “warm” and other colors as “cold.” The specific transition point from warm-to-cold is in the eye of the beholder, but as a general rule: reds, oranges, and yellows are perceived to be warm, and greens, blues, and purples are perceived to be cool.

This is important for us to remember for two reasons:

Warm colors tend to pop out towards us, and cool colors tend to recede. If you’re using one warm and one cool color, the warm one might feel slightly more dominant.
We can leverage color temperature to help imply relationships and differences in our data series. When using colors near each other in the color wheel, it’s good to keep them all cool or all warm; when using colors distant from one another, it’s helpful to have the tension of one cool and one warm, or a dominant warm and several cools.

Selecting a key color

Our first step in building a palette, then, is to select a key color. This color might be:

Our dominant brand color
A color that is prominent in our existing slides
A color found in an image that will appear near our chart
A color that evokes the right “feel” for the data, based on cultural associations

Regardless of how you select it, this key color will be used to denote the data points, or the data series, on which you feel it is the most important for your audience to focus.

All of the other colors we use will be based on where they are on the color wheel in relation to this key color, how many colors we intend to use, and what kind of relationship the rest of the data has to the data represented by the key color.

In the examples to follow, our key color is going to be orange.

In picking other colors for our palette, what should we be changing: the hue; the saturation; and/or the lightness?

That key color, whatever it is, will have a hue, a saturation, and a lightness. As we set out to select harmonious colors for our key color, we keep some of these values the same across each color we use. That consistency ties the palette together and makes our chart visually pleasing. Instead of changing the values of all three qualities for each color in our palette, ideally we are only changing one.

Which quality do we change? That depends on what we are using color to achieve. For the most part, with categorical data, we should change the hue, and with continuous data or values, we should change the saturation or the lightness. This is because we are more likely to perceive changes in saturation or lightness as having a quantitative component.

Palettes for comparing two things

If there’s only two things to compare, and one is more important than the other, then consider using your key color for the important one and gray for the less important one.

However, if there are two distinct things you want to highlight out of a field of, say, a dozen things; or if among dozens of elements, you want to distinguish subgroups that have particular qualities, then you’ll want to use two different focus colors.

Here are some color harmonies to consider.

For highlighting two series with no value judgment: analogous

An analogous harmony is very simple. We start from our key color, and find a color exactly one step to the left or right of it on our color wheel, at the same level of saturation. If you use neighboring colors on the wheel, neither will be more emphasized than the other. If your key color is your brand color, it might carry more weight with an internal audience, but the difference will be subtle.

For highlighting two series with a positive/negative connotation: complementary

A relatively well-known quotation from 20th-century artist Marc Chagall states that “colors are the friends of their neighbors, and the lovers of their opposites.” A key color can be supported well by the colors that are near to it on the color wheel, but much more strongly by the colors that are on the opposite side.

Complementary colors are direct opposites, and offer the strongest possible contrast. That makes them good for showing positive/negative distinctions. With complementary harmony, your key color (if it’s your brand’s main color) can be positive, and its complementary color can represent the negative.

It’s advisable not to use your brand’s main color to mean something negative, even if that’s how that color is commonly used. For example, to show profits and losses, black commonly represents profits (“in the black”) and red represents losses (“in the red”). So in this case, we might instead use:

red (our brand color) to mean gains, and blue (its complementary color) to mean losses; or
red for gains and gray for losses (to focus on our company’s success); or
gray for gains and blue for losses (to focus on where we need to improve).

For highlighting two series when one is of primary focus: near complementary

Whenever two colors come from opposite sides of the color wheel, you’ll have sufficient contrast to distinguish them without implying that they are related. A near-complementary harmony, instead of being 50% of the way around the color wheel, is 33% of the way around; our key color is at 1 o’clock, and our near-complements would be at either 5 o’clock or 9 o’clock.

Remembering that warm colors pop out more than cool colors, ideally your key color would be warm and your complementary color would be cool; if this is not the case, you can lessen the impact of your secondary color by decreasing its saturation slightly or changing its lightness to have less contrast with the background (usually, making it lighter).

Palettes for comparing three things

Choosing three colors that work well together is a significantly more challenging task than picking only two. Consider also that we want to pick colors that are visually pleasing but also imply the specific relationships among the different categories, or data points, that we want our audience to understand.

Here are some harmonies that can support a tricolor visualization.

For highlighting three series with no value judgment: analogous or triadic

With three different series—just like when there are only two series—analogous harmony works if you’re simply making categorical distinctions. Instead of selecting one neighboring color to your key color, you use both. In this case, the key color will have a stronger visual emphasis than its analogues, but only slightly.

You could also use triadic harmony, which includes three colors evenly-spaced around the color wheel. Triadic harmony has more contrast than analogous harmony, so it could be a better choice for presentations on big screens to large crowds. The downside is that it will not feel particularly elegant, and you lose the feeling of one color being the key color.

For highlighting one series against two other related series, or against two sub-components of the main series: split complementary

In this harmony, your key color sits alone on one side of the color wheel, and your two additional colors are on the opposite side, each one step away from the key color’s exact complement. This three-color palette emphasizes that the two secondary series are related to each other, but are distinct from the series represented by the key color. One sample use case for this color palette would be to show a cumulative series (for instance, “total sales”) in your key color, and then show each component of that series (“domestic sales” and “international sales”) as one of the two complementary colors.

In the image above, you can also see how the split complementary harmony still works if we start from a different key color and a different saturation level. All of the color harmonies we are discussing in this article are harmonies of hue.

Once you get comfortable with selecting a preferred harmony for your use case, you can experiment with adjusting the saturation or lightness of your secondary colors as well.

Palettes for comparing four things

The occasions where you truly need four distinct colors in a single visualization, hopefully, are rare. As mentioned above, color is best employed to focus attention, and if there are four unique colors in your visual, then it’s hard to say where people will focus. Nevertheless, those situations will arise from time to time, and these are the color harmonies that can help to subtly form associations in your audience’s mind about the relationships among your data series.

For one main series and its three components: analogous complementary

Analogous complementary harmonies include both of the key color’s split complements as well as its direct complement, for a total of four colors. Notice how the contrast makes the key color stand out against the complementary colors. This can help to ensure that the series you think is the important one, or the series that might represent the trend or the average, is the series that is most likely to stand out.

For two pairs of related series where one pair is dominant: double complementary

If you have four different data series, and they can be thought of as two groups of two series, then you might want to use double complementary harmony. For this palette you start with the key color, pick one of its two analogues, and then use the exact complements of those two colors. In this harmony, it helps if your key color and its analogue are both warm or both cool, and for the complements to be the opposite color temperature.

For categorically distinguishing four series of equal emphasis: rectangular or square

If you’re simply using color to make categorical distinctions across four series, with no one series necessarily more important than any other, then square or rectangular harmonies could be the right answer.

In rectangular harmony, you use a key color, a “near analogue” two steps away on the clock, and the complements of those two colors.
In square (or tetradic) harmony, you start from the key color and then use every third step as you go around the clock.

The rectangular harmony still retains a subtle hint that the four series might actually be two pairs of two series, but the square harmony places all four series on completely equal footing.

Palettes for showing quantitative value (color ramps)

As mentioned above, when using color to show quantities (as you might in a highlight table or a filled map), the differences in values are represented most often by changes in saturation and/or lightness.

For showing changes in value from zero to a maximum value: sequential color ramp

In this case, you would use a monochromatic (single-hued) palette, where the lowest value is represented by a color that matches (or nearly matches) the background color of your chart, and the highest value matches your key color, most likely with a 100% saturation and a 50% lightness. You could have a smooth color gradient, for more precision; or discrete color steps, which are easier for your audience to tell apart.

For showing changes in value through a range with a meaningful midpoint: diverging color ramp

A diverging color ramp is like two sequential ramps facing opposite directions, stitched together tail-to-tail at your data’s natural midpoint. If your data includes both negative and positive numbers, for instance, then you would choose a dichromatic (two-hued) palette, ideally with your key color (for positive) and its complementary color (for negative), and use a neutral gray color for the midpoint (where the tails are stitched together).

It’s important to note that the diverging color ramp’s midpoint is not the midpoint of your data range, but rather at the meaningful midpoint that is the threshold between positive and negative. (A data range from -10 to 90 should not have a color ramp that uses 40 as its midpoint. The color ramp, in fact, should run from -90 to 90, with a midpoint of 0.)

A few final thoughts and some resources

If you find yourself wanting to add more contrast between colors in your harmonies, you can adjust the saturation of the secondary colors down (making them paler), as long as you adjust all of the secondary colors’ saturations by the same amount.
If your primary brand color is cool, and the warm complementary colors seem like they’re overwhelming it, decreasing the saturation of those complementary colors can help to de-emphasize them further.
There are many drawing and painting applications that have integrated tools to generate palettes based on a key color and your chosen harmony: Adobe Illustrator and Procreate are two that I use regularly. Paletton is a free online tool that accomplishes the same thing and lets you output your chosen palette in a number. of different formats.
Make sure that the palettes you choose are colorblind safe. Around 9% of men have red-green colorblindness, and there are plenty of other men and women with color deficiencies in their vision. Use an online color-checker like Coblis to make sure your palettes are accessible to your audience.
If you happen to be doing cartography, and you are looking to select palettes that are going to work for mapping, ColorBrewer 2.0 is an excellent resource for that.
We have a conversation going in the SWD community all about resources to use when making decisions about color. Get inspiration from what others have shared there, and add your own favorite resources as well!

Now it’s your turn!

Now that you’ve learned all about color, why not apply these lessons to this exercise in the SWD Community?

picking the right colors