geom-boxplot

Boxplot layer: draws a Tukey box, whiskers, and outlier points per group.

Defaults to stat = "boxplot", so the layer accepts raw observations and computes the five-number summary per group internally. The default position = "identity" keeps groups at their categorical x slot; switch to "dodge" if you want side-by-side boxes per fill level.

Usage

geom-boxplot(
  mapping: none,
  data: none,
  width: 0.6,
  colour: auto,
  fill: auto,
  stroke: 0.6pt,
  alpha: auto,
  outlier-size: 1.8pt,
  outlier-colour: auto,
  whisker-cap: 0.5,
  stat: "boxplot",
  position: "identity",
  inherit-aes: true,
)

Parameters

Parameter	Default	Description
mapping	none	Layer-specific aesthetic mapping built with aes. Falls back to the plot mapping when none.
data	none	Layer-specific dataset. Falls back to the plot data when none.
width	0.6	Box width in x data units. For discrete x this is also a data-unit width.
colour	auto	Stroke colour for the box, median, and whiskers. auto falls back to the theme ink only when neither colour nor fill is set.
fill	auto	Box fill colour. auto resolves via the fill scale or a neutral default.
stroke	0.6pt	Stroke thickness for the box outline and whiskers.
alpha	auto	Box opacity in [0, 1].
outlier-size	1.8pt	Marker size for outlier points.
outlier-colour	auto	Marker colour for outlier points. auto follows the box stroke colour.
whisker-cap	0.5	Cap length at the whisker ends as a fraction of width.
stat	"boxplot"	Statistical transform name. "boxplot" by default.
position	"identity"	Position adjustment name. "identity" by default.
inherit-aes	true	Whether to merge the plot-level mapping into this layer’s mapping.

Returns

Layer dictionary consumed by plot.

Examples

Five-number summary computed per category from raw observations.

#let d = ()
#for grp in ("a", "b", "c") {
  for i in range(20) {
    d.push((grp: grp, y: calc.sin(i) + i / 10))
  }
}
#plot(
  data: d,
  mapping: aes(x: "grp", y: "y"),
  layers: (geom-boxplot(),),
  width: 10cm,
  height: 6cm,
)

Add a second mapping (fill) and switch position to "dodge" to compare distributions side by side per group.

#let d = ()
#for grp in ("a", "b", "c") {
  for k in ("x", "y") {
    for i in range(20) {
      d.push((grp: grp, k: k, y: calc.sin(i) + i / 10 + (if k == "y" { 0.7 } else { 0 })))
    }
  }
}
#plot(
  data: d,
  mapping: aes(x: "grp", y: "y", fill: "k"),
  layers: (geom-boxplot(position: "dodge"),),
  width: 10cm,
  height: 6cm,
)

See also

stat-boxplot, geom-col.