From pg. 67 of Edward Tufte’s Envisioning Information:
At the heart of quantitative reasoning is a single question: Compared to what? Small multiple designs, multivariate and data bountiful, answer directly by visually enforcing comparisons of changes, of the differences among objects, of the scope of alternatives. For a wide range of problems in data presentation, small multiples are the best design solution.
What are small multiples? Essentially, a small multiple is a series of displays with the same design structure repeated for all the images, arranged in a grid. That means each graph in the series should be the same size and shape, with the same scale, differing only in the data they display.
What are the advantages of using small multiples? On page 29 of Envisioning Information, Tufte says, “An economy of perception results; once viewers decode and comprehend the design for one slice of data, they have familiar access to data in the other slices. As our eye moves from one image to the next, this constancy of design allows viewers to focus on changes in information rather than changes in graphical composition. A steady canvas makes for a clearer pictures.” (If you want to learn more about small multiples, all of chapter four is dedicated to them.)
Unfortunately, most of Tufte’s examples in Envisioning Information, e.g. the proper formation of capital letters, light signals for a train, or Saturn’s orbit, while instructive, are a bit of a stretch to apply to common BI situations. Enter Stephen Few, who always manages to apply Tuftean principles in a way that you can use them at work. From page 159 of Information Dashboard Design:
Concerning their efficiency, a small multiple offers another advantage over a series of individual graphs: the title, legend, and other metadata need to be printed only once to represent the series.
Here, Few uses small multiples to introduce another dimension to the standard grouped bar chart:
The last example from Stephen Few I’ll mention is that small multiples can be used as “visual crosstabs”. (Of course, it is helpful to have the supporting information available, too, if possible.)
Improve Your Vision (PDF)
A white paper worth reading that covers some of the principles involved when employing small multiples is Three Blind Men and an Elephant: The Power of Faceted Analytical Displays (PDF).
Looking at other examples, the written-in-stone-reliable (cough) Wikipedia‘s sample image in the entry for small multiple is not a small multiple, due to the different metrics and scale of vertical axis in each chart.
Another small multiple fail is way back in one of the first posts on this blog, The Trilogy Meter. The problem there is that the graphs are arbitrarily arranged, while they should be in order of magnitude from greatest overall to worst overall trilogy.
My image sample from the QlikView 9.0 Beta – QlikView 9 being the first release to support trellis charts – may not have been a great example of when to use small multiples. They should not be frivolously used in place of every multi-series line graph, containing the same information in merely five times the space. That said, it may be of value if you see small multiples as an alternative to using a list box to toggle through slices of a dimension, due to the way the human memory works, as discussed in the “single context” section of my post about facilitating comparison.