Get Visual With Small Multiples


Small multiples are series of small, similar graphics or charts that, when used appropriately, make it easier for the reader to compare different data series at a glance. Lots of people love them, including Ben Jones from Tableau, Ann Emery from Innovation Network, and Purna Duggirala from Chandoo.org.

You can create small multiples in all sorts of software packages, including Stata (through the twoway scatter or graph combine commands), R (the lattice library), or SAS (the proc greplay procedure). Small multiples can also be created in Excel, though that's actually a little trickier, because you have to resize the charts and line them up manually. An external add-in like some of the ones Jon Peltier discusses on his website can make those tasks slightly easier.

There is another way to create small multiples in Excel without having to create and manage multiple graphics manually. Here, I'll walk you through that process with three-column charts. The method is an example of how to extend Excel's capabilities, especially by using scatterplots.

Click the image below for a step-by-step guide on how to change a standard column chart to one that looks as if you've created separate charts.

In some cases, encoding data to create specialized Excel charts is preferable to creating separate charts and piecing them together manually. Using data to encode labels, extra lines, spaces, or other features can produce more flexible charts that you can modify and update.

Of course, these extensions will only get you so far. Programming languages like SAS, Stata, and R will let you do much more. But the extensions are a good start.

Do you love small multiples? How do you use them? Share on the message board below.

Related posts:

Jonathan Schwabish, US Economist & Data Visualization Creator

Jonathan Schwabish is an economist with the US federal government and creator of policy-relevant data visualizations. He has conducted research on inequality, immigration, retirement security, data measurement, food stamps, and other aspects of public policy in the United States. His work has been published in the Journal of Human Resources, the National Tax Journal, and elsewhere. He is also a data visualization creator and has made designs for a variety of subjects that range from food stamps to healthcare to education. His visualization work has been featured on the visualizing.org and visual.ly websites. Jonathan has spoken at numerous government agencies and policy institutions about data visualization strategies and best-practices. His website, policyviz.com, contains details about his one-day workshops on visualizing and presenting data for people in public policy. He earned his PhD in economics from Syracuse University and his undergraduate degree in economics from the University of Wisconsin at Madison.

Get Visual With Small Multiples

When used appropriately, small multiples make it easier to compare different data series at a glance.

To Label or Not to Label? That Is the Question

Sometimes showing data values in a chart interferes with the visual representation.


PROC SGPANEL is useful here
  • 10/3/2013 11:20:00 AM
NO RATINGS

You mentioned the old PROC GREPLAY in SAS, so I wanted to add that for many standard statistical graphics you can generate "small multiples" by using the more modern ODS graphics through PROC SGPANEL. Some examples are in the SAS documentation.

Re: easy as 1, 2, 3
  • 8/15/2013 11:48:30 PM
NO RATINGS

@Johnathan: Thank you for sharing the links and your thoughts on this. I will be having some questions for you surely. I will direct them to you after going through the links.  

Re: easy as 1, 2, 3
  • 8/15/2013 9:09:30 PM
NO RATINGS

Another great example. Thanks!

Re: easy as 1, 2, 3
  • 8/15/2013 8:20:57 PM
NO RATINGS

@Beth, the method I've described here really only works for small multiples that run horizontally (for column charts; vertically for bar charts). For more charts--especially for more rows of charts--it becomes more complicated. In those cases, I would suggest a programming language like SAS, Stata, or R, or creating Excel charts that you then align to one another. A great example can be found at Ann Emery's latest post: http://emeryevaluation.com/2013/07/10/small-multiples/ and http://emeryevaluation.com/2013/07/27/dataviz-challenge-4-the-answers/

easy as 1, 2, 3
  • 8/15/2013 7:28:33 PM
NO RATINGS

So Jonathan, your directions seem straightforward enough.But if somebody is going to get hung up creating small multiples using this method where would it be?

Re: What to know?
  • 8/15/2013 2:42:07 PM
NO RATINGS

@urbie4 makes some very good points here, @Noreen. For me, the difference between the NYTimes maps and the one you posted is that the NYTimes maps present a binary property and thus it's much easier to visualize the patterns. Each of the maps in this graphic, however, are choropleth maps (thematic map in which the states are differentially colored based on the encoded data), which makes it difficult to discern the pattern. A potentially better graphic would have been to just color the states that exceed some threshold (for example, if the share of people in the state in that cell exceeds 25%, then it gets a dark green color).

Re: What to know?
  • 8/15/2013 2:34:20 PM
NO RATINGS

@urbie4, I agree. This is way too muddied. The designers were trying to do too much here.

Re: What to know?
  • 8/15/2013 2:24:09 PM
NO RATINGS

Jonathan, I checked out the small multiple maps you pointed out at the NYTimes. It certainly chews up a lot of space, but you're right, it's a good way to show a high-level, at-a-glance look at US drought over the years. 

Re: What to know?
  • 8/15/2013 12:04:15 PM
NO RATINGS

@Noreen To me, that graphic has way too much in it for me to absorb.  With 40 maps of the Continental US, I can't see the forest for the trees.  Maybe if it used fewer income divisions, that might help.  The patterns for White Catholics and White Evangelicals look fairly similar, so maybe lump them together.  (If you're going to talk about "non-evang. Protestants," though, that's a touchy subject -- I'm an Episcopalian, and we have a lot of crossover/cooperation with the Evangelical Lutheran Church in America, which is not "evangelical" in the speaking-in-tongues sense. My church used to have a substitute pastor who was from the ELCA, and he was pretty liberal, almost like an Episcopalian! :-D  Which is itself a gross generalization, of course, given that Karl Rove is one. But I digress.)  The White non-evang. Protestants don't look all that different from the no-religion types, on the map -- maybe you could put them together, as well.  But do something to reduce the number of maps!  As for the number of color gradations, it's been 30 years since I took cartography, but I think they said the eye couldn't really distinguish more than around 5-6 degrees of gray scale.

Re: What to know?
  • 8/15/2013 11:06:55 AM
NO RATINGS

Do you like this?

Page 1 / 2   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +