Small Multiples
A grid of identical mini-charts, each showing a different slice of the data — the same axes, the same scale, so differences between panels jump out instantly.
// 01 — The chart
What it looks like
North
South
East
West
Central
Overseas
Six small line charts — one per region — using the same Y-axis scale. The Central region shows the steepest decline.
// 02 — Definition
What are small multiples?
Small multiples (also called a trellis chart, lattice chart, or panel chart) is a technique where the same chart type is repeated in a grid, each showing a different slice of the data. The key constraint: all panels use identical axes, scales, and visual encoding so they can be compared at a glance.
The term was popularized by Edward Tufte, who called them “the best design solution for a wide range of problems in data display.” The idea is powerful: instead of cramming 6 lines onto one chart (where they overlap and create spaghetti), you give each line its own panel. Patterns and outliers become immediately visible.
Small multiples work with any base chart type: line charts, bar charts, scatter plots, maps — anything. The power comes from the consistent structure that lets the eye detect differences rapidly.
Edward Tufte: “At the heart of quantitative reasoning is a single question: Compared to what? Small multiples, by presenting a series of views in a consistent framework, answer this question directly.”
// 03 — Anatomy
Parts of a small multiples layout
Panel A
Panel B
Panel C
// 04 — Usage
When to use it — and when not to
- A single chart would have too many overlapping series (spaghetti lines)
- You want to compare the same pattern across categories, regions, or time periods
- Each subset has the same structure (same X/Y dimensions, same scale)
- The audience needs to spot outliers or anomalies by scanning a grid
- You have 4–20 subsets to compare simultaneously
- You have more than ~20 panels — they become too small to read
- The subsets have very different scales — a shared scale would squash some panels
- The primary comparison is between data points within a single panel (use a regular chart)
- There are only 2 subsets — a single overlay chart is simpler
- The audience is unfamiliar with the base chart type used in each panel
// 05 — Reading guide
How to read small multiples
Read one panel fully first
Pick any panel and understand its chart type, axes, and scale. This panel is your template for reading all the others.
Confirm shared scales
Check that all panels use the same axis ranges. If they do, you can compare shapes directly across panels.
Scan for the outlier
Glance across all panels — which one looks different? This is where the story usually lives.
Compare shapes, not values
The power of small multiples is pattern comparison: rising vs flat vs declining, smooth vs spiky.
Read panel labels
Each panel is labelled with the subset it represents. Use these labels to interpret the pattern you noticed.
// 06 — Data format
What your data should look like
| Column | Type | Description |
|---|---|---|
| Facet | String | The variable that determines which panel each row appears in (e.g., region) |
| X | String / Number | The shared X-axis variable (e.g., month) |
| Y | Number | The measured value |
North, Jan, 42
North, Feb, 38
South, Jan, 55
South, Feb, 50
...
// 07 — Construction
How to build small multiples
Choose the faceting variable
Decide what splits the data into panels: region, product category, year, etc.
Pick a shared chart type
All panels must use the same chart type (line, bar, scatter) for consistency.
Lock the axes
Set a single Y-axis range across all panels. Optionally share the X-axis too. This is critical for fair comparison.
Arrange in a grid
Lay out panels in a logical order (alphabetical, geographical, ranked). Use a consistent grid like 3×2 or 4×3.
Label each panel
Add a clear title to each panel stating the subset (e.g., 'North', 'Q1 2025').
// 08 — Pitfalls
Common mistakes
Using different Y-axis scales per panel
This is the cardinal sin of small multiples. Different scales destroy comparability. Lock the range across all panels.
Too many panels
Beyond ~20, panels become too small and the grid too large to scan. Aggregate or filter to a meaningful subset.
Inconsistent panel sizing
All panels must be the same size. If one panel is bigger, it draws disproportionate attention.
No panel labels
Without labels, readers can't identify which subset each panel represents. Always title every panel.
// 09 — In the wild
Real-world examples
The New York Times election coverage
Grid of state-level trend charts showing polling averages, enabling readers to quickly spot battleground states.
COVID-19 dashboards
Small multiples of case curves by country became a staple of pandemic reporting, revealing which nations flattened the curve first.
Financial earnings reports
Analysts compare revenue trends across business segments using identical quarter-by-quarter line charts in a grid.
// 10 — Quick reference
Key facts
| Also known as | Trellis chart, lattice chart, panel chart, faceted chart |
| Popularized by | Edward Tufte (1983) |
| Base chart type | Any (line, bar, scatter, map, etc.) |
| Ideal panel count | 4 – 20 |
| Cardinal rule | All panels must share the same axes and scales |
// 11 — Accessibility
Accessibility notes
Provide a summary table
A single table with all facets and values is far easier for screen readers than navigating multiple chart SVGs.
Label each panel with ARIA
Each panel container should have aria-label="Sales trend for North region" for assistive technology.
Maintain sufficient panel size
Panels that are too small (below ~100px) become unreadable for low-vision users. Set a minimum size.
Use consistent colour
The same colour should mean the same thing across all panels. Don't rotate palettes between facets.
// 12 — Variations
Variations
Sparkline grid
Ultra-minimal panels — no axes, just the shape — for the most compact comparison
Faceted maps
Each panel is a choropleth map showing a different time period or variable
Linked small multiples
Interactive: hovering on one panel highlights the same point in all panels
Free-scale facets
Each panel has its own scale — trades comparability for detail within each facet
// 13 — FAQs
Frequently asked questions
What is a small multiples?+
Small multiples (also called a trellis chart, lattice chart, or panel chart) is a technique where the same chart type is repeated in a grid, each showing a different slice of the data. The key constraint: all panels use identical axes, scales, and visual encoding so they can be compared at a glance.
When should you use a small multiples?+
Use a small multiples when a single chart would have too many overlapping series (spaghetti lines). It also works well when you want to compare the same pattern across categories, regions, or time periods, and when each subset has the same structure (same X/Y dimensions, same scale).
When should you avoid a small multiples?+
Avoid a small multiples when you have more than ~20 panels — they become too small to read. It is also a poor fit when the subsets have very different scales — a shared scale would squash some panels, or when the primary comparison is between data points within a single panel (use a regular chart).
Are small multipless accessible to screen readers?+
Yes — a small multiple can be made accessible to screen readers by pairing it with a clear text summary of the key insight, ensuring color choices meet WCAG contrast guidelines, adding descriptive alt text or aria-label to the SVG, and offering the underlying data as an HTML table fallback for assistive technologies.
Is a small multiples suitable for dashboards?+
Yes — a small multiples can work well in dashboards as long as the panel is large enough for readers to perceive the encoded values, has a clear title, and includes the legend or axis labels needed to interpret it.
What category of chart is a small multiples?+
Small Multiples belongs to the Comparison family of charts. Charts in that family are designed to answer the same kind of question, so they often work as alternatives when one doesn't quite fit your data.