Create rug plots representing distribution of the raw data

The companion to a smooth plot: it shows where the data actually is, so a bend in a smooth can be read against how much evidence sits under it.

Usage

plotRugs(
  dat,
  var,
  type = c("histogram", "density"),
  transform = c("none", "log", "log10", "sqrt"),
  bins = 30,
  fill = "grey35",
  group = NULL,
  palette = NULL
)

Arguments

dat

Raw data

var

Variable to plot

type

Optional parameter indicating type of plot; default is histogram

transform

Optional parameter indicating how to transform the variable, if applicable

bins

Number of histogram bins. Set explicitly rather than left to geom_histogram()'s default, which is the same 30 but emits a message about it on every plot. Ignored when type is "density".

fill

Fill colour for the rug. Deliberately a neutral grey: the rug reports where the data is, and should not compete with the effect curve below it for attention. Ignored when group splits the rug, which colours it by level instead.

group

Optional name of a factor column in dat to split the rug by, as a string. Under a factor-smooth interaction each curve is fitted to one level's data only, so an undivided rug reports evidence that does not belong to the curve the reader is looking at. Split, each level's share of the data is drawn in that level's own colour.

palette

Colours for the split, as a vector named by level – the names are what tie a band of rug to its curve, so an unnamed vector is assigned in level order and a NULL leaves the scale to ggplot2. Levels the palette does not name are dropped from the rug rather than drawn in a colour that belongs to another curve.

Value

The rug plot from dat for var

See also

plotEffects(), which stacks this above an effect curve for you.

Other effect plots: combinePlots(), comparePlots(), plotEffects(), plotSmooths()

Examples

plotRugs(iris, "Sepal.Length")

plotRugs(iris, "Sepal.Length", type = "density")

plotRugs(mtcars, "disp", transform = "log10", bins = 15)


# Split by a factor, in the palette the effect curves use
plotRugs(iris, "Sepal.Length", group = "Species",
         palette = fancyfx_palette(3))