I’ve been thinking about how to make my chart legends more engaging. I prefer to directly label my data rather than using legends that are separated from the data, but when I need a legend—say, on a map or a denser bar chart or scatterplot—are there ways to really attract the reader’s attention?

“Clever legends” are those that take an interesting approach to defining the data elements in the chart. Instead of a standard set of squares or circles or lines with labels, ‘clever legends’ use a unique charting device or design to engage us while also simply providing us with the data definitions. After recently completing my read of Nigel Holmes’s wonderful book, Joyful Infographics, I might consider some of these examples “joyful legends.” Perhaps my favorite clever legend is this famous bubble chart from Hans Rosling. Notice the map-as-legend in the bottom-right part of the image where color shows each region of the world, which is used in the bubble chart itself.

Gapminder bubble chart that shows the positive relationship between GDP per capita and life expectancy.
Source: Gapminder

It’s worth noting that I’m distinguishing these kinds of clever legends from toehr legend types that may also be useful and engaging, but, say, use icons or shapes in the actual graph. As an example, the legend in this graph—from a really nice scrollytelling piece by Karim Douieb—defines the values of each of the icons.

Icon chart with a series of black and white silhouettes of people representing those born in Belgium and those not born in Belgium. The chart is about 50-50.
Source: Karim Douieb

I also don’t mean legends that have a smaller version of the graphic that are used to directly explain the visualization. For example, this tile grid map from the US Geological Survey has an area chart of streamflow for every state in the country in July 2021. The legend over to the side is a larger version of each tile with a legend built in. A good legend to be sure, but not the class of ‘clever legends’ I have in mind.

Tile grid map of the United States showing the streamflow in each state in July 2021.
Source: USGS

Clever legends don’t need to be in clever or bespoke graphs. The USDA published this stacked bar chart this week with color icons in the legend. The color+icon approach is a little fun and still easy to read. (The stacked chart may not be the best chart type for these data, but that’s another story.)

Stacked bar chart showing the percent of manure applied across 7 crop types from 5 sources of manure: dairy cow, beef cow, hogs, other, and poultry.
Source: USDA

Similarly, this map of world avocado production breaks the world down into four groups and the colors for each are encoded in the–wait for it, because this is amazing–“Avo-key-do” at the bottom left (I’ve enlarged it just below). Simple pointers to the parts of the little avocado icon use colors to differentiate the four groups. Now, one could argue that this should a sequential color palette and the brown (pit) should be the lightest green, but I think I’d argue that setting the zero category off from the others outside the green sequential palette works just as well.

World map of avocado production and estimated exports.
Source: Reddit
Zoom in of the key to the avocado map. The legend is a drawing of an avocado with the title Avo-key-do.

The legend in this bivariate choropleth map from FiveThirtyEight is not necessarily clever in the way I’m defining it, but does a great job explaining how to read the map. The three small multiple versions at the top help you understand each dimension of the map and is then paired with full legend below the menu of state names.

Here’s a good, simple example of a clever legend. Will Chase from Axios created this tile grid map where the houses are sized by the share of renters who are not current on their rent. In addition to using color in the legend, this legend has a little house icon, scaled to the value. It’s a little play on the icon-as-legend technique we saw earlier.

I’m a big fan of using bar-charts-as-legends in maps. This is a good example from my colleagues at the Urban Institute where the four groups shown in the map are presented as a bar chart in the legend. We don’t even have to examine the map to see that a majority of counties fall in the 10.1% to 20% bucket.

Finally, the little gauge-as-legend in this cool piece from Possible. on noise in the New York City area also strikes me as clever. Here, because we often see measures of sound on a radial gauge-type visual, it works really well here. (Side note: this piece was inspired by work from Karim Douieb showing noise in Brussels. I don’t know Karim and it’s completely coincidental that two of their pieces are showing up in this post—well, maybe not totally coincidental as I’m getting a sense they try to play with clever legends. Kudos, Karim.)

Map of the New York City metro region with light to dark pink lines showing the levels of exposure to noise.
Source: Possible

I don’t have a ton of examples of clever legends, which is probably whey I think they’re clever. I’m on the lookout for more and to consider ways that I might incorporate these kinds of things into my own work in the future. If you see an example of a clever legend, be sure to let me know!

Did you like this post? You could have seen it a few weeks ago in my free bi-monthly newsletter, now on Substack!