Skip to contents

Draw state events as horizontal bars

Source: R/geom_event_state.R
geom_event_state.Rd

A ggplot2 layer that renders state (durative) events as rectangles spanning each event's start -> stop on the x axis, with one row per label on the y axis. When the input has a type column (e.g. an anicore::anievent()), rows with type != "state" are dropped.

Usage

geom_event_state(
  mapping = NULL,
  data = NULL,
  ...,
  height = 0.7,
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

Arguments

mapping

Aesthetic mapping. Merged on top of the defaults, so you only need to supply the aesthetics you want to override.

data

A data frame (typically an anievent). If NULL, inherits from the parent ggplot() call. Either way, rows with type != "state" are dropped from this layer.

height

Vertical extent of each bout bar, in y-axis units (1 = the spacing between adjacent label rows). Default 0.7.

na.rm, show.legend, inherit.aes, ...

Passed to ggplot2::layer().

Value

A ggplot2 layer.

Details

The default mapping is only aes(xmin = start, xmax = stop) — every other aesthetic is up to the caller. Common patterns:

  • aes(fill = label) — colour each bout's interior by behaviour.

  • aes(colour = label) — outline each bout (keeps sub-pixel-wide bouts visible at any zoom).

  • aes(y = label) (or y = channel, y = subject, ...) — spread bouts across rows. With no y mapping all bouts collapse onto a single line.

  • + ggplot2::facet_wrap(~ channel, scales = "free_y", ncol = 1) — split channels into their own panels.

Opinionated wrappers (themes, colour scales, sensible defaults) are intended to live one layer up; this geom is deliberately bare.

This is a thin wrapper around ggplot2::GeomRect: a subclass adds a setup_data step that derives ymin / ymax from y + height, so a discrete y scale keeps the label rows ranked correctly (including under per-panel scales = "free_y").

Examples

library(ggplot2)
af <- anicore::example_aniframe(n_obs = 6, n_individuals = 1, n_keypoints = 1)
af$behaviour <- rep(c("walk", "rest"), each = 3)
ev <- anicore::to_anievent(anicore::set_variables_event(af, state = "behaviour"))
ggplot(ev, aes(xmin = start, xmax = stop, y = channel, fill = label)) +
  geom_event_state()