Key Takeaways — Chapter 8: Layout, Composition, and Small Multiples

1. Composition Is a Distinct Design Skill

A collection of well-designed individual charts is not automatically a well-composed figure. Composition — the arrangement of multiple charts on a page — is its own skill, with its own rules, and it does work that no single chart can do: enabling comparison, establishing hierarchy, guiding reading order, and creating visual unity. A figure with five perfect charts arranged badly is still a bad figure. Composition is worth deliberate attention, not just a default layout.

2. The Four Jobs of Composition

Every multi-chart figure should be evaluated against four criteria: (1) Enable comparison — do related charts share axes, scales, and placement that allow visual comparison? (2) Establish hierarchy — is the most important chart visibly larger or more prominent than the supporting charts? (3) Guide reading order — does the sequence of panels match the logical order of the argument? (4) Create visual unity — do all the panels feel like parts of one figure, through consistent typography, color, and alignment?

3. Gestalt Principles Apply at the Layout Level

The same perceptual principles that govern single charts — proximity, similarity, alignment (continuity), enclosure — govern how the reader perceives multi-chart figures. Related panels should be near each other (proximity). Panels that belong together should share visual characteristics (similarity). Panel edges should align to a visible grid (continuity). High-level groupings can be signaled by shared boundaries (enclosure). These are not metaphors — the visual system really applies the same grouping rules at the figure level that it applies within a single chart.

4. Small Multiples Are the Most Powerful Composition Technique

A small multiple is a set of charts that share the same visual encoding — same chart type, same scales, same design — differing only in which slice of the data they show. Tufte called them "the best design principle," and while the claim is overstated, small multiples do solve more visualization problems than any other single technique. They work by enforcing comparison: because every panel uses the same grammar, the reader can compare across panels pre-attentively, without the cognitive cost of re-learning each chart.

5. Consistency Is the Precondition for Comparison

The threshold concept: you can only compare panels that share the same visual encoding, the same scales, and the same baseline. Violate the consistency and the small multiple becomes a gallery of incomparable pictures. This is why mixing chart types in a small-multiple layout defeats the purpose: the reader cannot compare across different chart grammars. The power of small multiples is not the multiplication of charts — it is the forced comparison that shared encoding enables.

6. Shared Axes Enable Comparison; Free Axes Enable Pattern Visibility

When designing small multiples, the biggest decision is whether to share the y-axis across panels or let each panel have its own free axis. Shared axes let the reader compare absolute magnitudes across panels; free axes let the reader see the shape of the pattern in each panel regardless of its absolute level. Neither is always right. The choice depends on which comparison matters, and the chart should label the axis choice clearly so the reader knows what kind of comparison is possible.

7. Aspect Ratio Matches Chart Type and Data

Wide aspect ratios work best for time-series charts, because the time dimension needs horizontal room and because Cleveland's "banking to 45 degrees" rule suggests matching the average line slope to 45 degrees. Square aspect ratios work best for scatter plots, because both axes are equally important. Tall aspect ratios work best for horizontal bar charts with many categories, because each category needs its own row. Choosing the right aspect ratio is a small decision with a disproportionate effect on legibility.

8. Reading Order Follows the Z-Pattern

Western readers scan multi-panel figures in a Z-pattern: top-left first, then top-right, then bottom-left, then bottom-right. The most important panel (the hero) should be placed in the top-left, where the reader's freshest attention lands. Panels should be ordered so that the sequence of the reading matches the sequence of the argument — context first, then the main finding, then implications. A layout that fights the Z-pattern forces the reader to work against their natural scanning pattern.

9. Dashboards Are Not Small Multiples

A dashboard shows different chart types for different aspects of a system; a small multiple shows the same chart type for different slices of the data. The two have different design rules. Small multiples depend on consistency across panels. Dashboards depend on creating unity from diverse elements: consistent typography, a single accent color, consistent whitespace, aligned edges. A dashboard designed with small-multiple rules is flat and lacks hierarchy; a small multiple designed with dashboard rules is inconsistent and lacks comparability. Know which problem you are solving.

10. Composition Amplifies But Does Not Rescue

Composition is a distinct skill, but it is not the answer to every problem. Before investing in compositional polish, apply the Chapter 6 and Chapter 7 disciplines to each individual chart: declutter it, give it an action title, add annotations, format the axes. Each chart should stand on its own as a decluttered, self-explanatory chart before you try to arrange multiple charts together. Composition amplifies what is already there — but a beautifully laid-out figure of bad charts is still a figure of bad charts.