U2.06.4Task-chosen reference values for bubble-size legendsdesignresearch

Bubble size needs a reference-bubble legend

Aliases: size legend · reference bubbles · bubble scale

What it is

A size legend labels reference circles with values and units so readers can interpret the bubble-area mapping. “Needs” in the title means that the third variable needs an available scale or value route, not that every view must contain a fixed number of circles. Direct labels, an accessible table, or interactive readouts can carry part of the function, but the static, exported, and unoperated states still need self-contained reference.

Why it happens

Area has no ticked axis. Readers compare plotted circles with labeled references to recover data units from relative size. The number and placement of reference values should cover the data range and decision thresholds, not mechanically select minimum, median, and maximum. Legend and plot bubbles must share the same area function, size range, and transformation; equal diameter steps in the legend paired with proportional area in the plot would make the reference itself misleading.

Studying it

Compare no legend, candidate reference sets, direct labels, and details on demand for magnitude, threshold, ratio, lookup-time, and error tasks. Manipulate distribution, skew, extremes, legend layout, and delivered size to see whether references cover the comparisons readers actually make. Interactive studies should include first load, keyboard focus, touch, screenshots, and offline export; success after hover alone does not establish that a legend is redundant.

Where it stops holding

Too many references consume space; too few, or extremes alone, may leave the middle hard to estimate. There is no universally best count. Heavy skew and logarithmic transforms require interpretable references and explicit original units. Direct labels work for selected marks but cannot densely annotate every bubble. The association among circles, labels, and plot must survive zoom, high contrast, and screen-reader output.

Applying it

  • Choose a small set of discriminating references from the data range, distribution, and task thresholds; state variable, unit, time scope, and transformation.
  • Reuse one size function and rendering component in legend and plot, with automated checks of rendered area for known values.
  • Add direct labels or exact details for priority bubbles, and expose values and states in the legend and every focusable mark's accessible name.
  • Test magnitude, threshold, and ratio tasks. If legend travel remains costly, reposition it, label directly, or add a coordinated detail view.

Related

  • Same group: U2.06.1 Bubbles carry a third variable on the least precise common channel · U2.06.2 Map values to bubble area, not radius · U2.06.3 Big bubbles occlude small ones; draw in size order and lower opacity · U2.06.5 Precise third-variable comparison belongs in facets or a separate chart
  • Adjacent: U3.05.3 Direct labelling can replace gridlines and legends · U1.11.2 Encoding that jumps between charts forces readers to relearn the legend
  • Search terms: size legend · reference bubble · legend lookup · direct labeling

Cards in the same group

Quick Actions

Share

Share this page

ios_share

https://hci.top/en/handbook/U2.06.4