Skip to contents

Circular density statistic

Source: R/stat-circular-density.R
stat_circular_density.Rd

Estimates a smooth circular density using a von Mises kernel. The density wraps around the origin, avoiding the boundary artifacts of a linear kernel density estimate. When bw is not supplied, the concentration is chosen from a simple resultant-length heuristic; it should be treated as an exploratory smoothing choice rather than an inferential bandwidth selector.

Usage

stat_circular_density(
  mapping = NULL,
  data = NULL,
  geom = "line",
  position = "identity",
  ...,
  method = c("kernel", "vonmises"),
  bw = NULL,
  adjust = 1,
  n = 512,
  axial = FALSE,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

mapping, data, geom, position, show.legend, inherit.aes

Standard ggplot2 layer arguments.

...

Additional arguments passed to the layer.

method

Density method. Currently "kernel" and "vonmises" both use a von Mises kernel estimator.

bw

Optional circular bandwidth. It is interpreted as 1 / sqrt(kappa).

adjust

Multiplicative adjustment applied to bw or to the automatic bandwidth scale.

n

Number of grid points.

axial

Should the data be treated as axial, modulo pi?

na.rm

Should missing values be silently removed?

Value

A ggplot2 layer. Computed variables are x, density, scaled, count, n, bw and kappa.

See also

Other circular density layers: geom_circular_density()

Examples

ggplot2::ggplot(wind_directions, ggplot2::aes(x = direction)) +
  stat_circular_density()