Density-shape marks

Marks that draw a variable's distribution as a filled shape. mark_violin() draws a mirrored kernel density of y per categorical x (like a boxplot's footprint); mark_ridgeline() draws a kernel density of x per categorical y as overlapping ridges. mark_dotplot() bins x and stacks one dot per observation.

Usage

mark_violin(plot, ..., adjust = 1, blend = NULL, sketch = NULL, data = NULL)

mark_ridgeline(
  plot,
  ...,
  adjust = 1,
  scale = 1.4,
  blend = NULL,
  sketch = NULL,
  data = NULL
)

mark_dotplot(plot, ..., binwidth = NULL, blend = NULL, data = NULL)

Arguments

plot

A PlotSpec.

...

Encodings. mark_violin() needs categorical x and numeric y; mark_ridgeline() needs numeric x and categorical y; mark_dotplot() needs x. A mapped color/fill sets the shape fill.

adjust

Kernel-density bandwidth multiplier (violin/ridgeline).

blend, sketch, data

Standard layer arguments (see mark_point()).

scale

Ridge height as a multiple of the row band (ridgeline; default 1.4, so adjacent ridges overlap slightly).

binwidth

Dot-plot bin width, or NULL to use ~1/30 of the data range.

Value

The modified PlotSpec.

Examples

df <- data.frame(g = rep(letters[1:3], each = 50), v = rnorm(150))
vplot(df) |> mark_violin(x = g, y = v)

vplot(df) |> mark_ridgeline(x = v, y = g)

vplot(df) |> mark_dotplot(x = v)