Many-plot helpers: ridgelines, ECDFs, violins, hexagonal bins and concentration curves

Base-graphics plots in the house style for a numeric variable split by a group, or two numeric variables:

plotrdg()

a ridgeline: one kernel density per group, stacked with a small overlap.

plotecdf()

the empirical distribution function of each group, overlaid.

plotvln()

a violin per group: the mirrored density, the median as a point and the interquartile range as a bar.

plothex()

the two-dimensional density of x and y in hexagonal bins, shaded by count.

plotgain()

cumulative-gain (concentration) curves: the Lorenz curve of each group (lorenzc()) with its Gini coefficient in the key.

Missing and non-finite values are dropped. A group of fewer than two values draws no density.

Usage

plotrdg(x, grp = NULL, bw = "nrd0", ..., main = NULL, col = NULL, pal = NULL)

plotecdf(x, grp = NULL, ..., main = NULL, col = NULL, pal = NULL)

plotvln(x, grp = NULL, bw = "nrd0", ..., main = NULL, col = NULL, pal = NULL)

plothex(x, y, bins = 30L, ..., main = NULL, col = NULL, pal = NULL)

plotgain(x, grp = NULL, ..., main = NULL, col = NULL, pal = NULL)

Arguments

x

Numeric vector to plot, one value per observation.

grp

Optional vector of groups, one per value of x.

bw

Bandwidth passed to stats::density(): one positive number or the name of a rule ("nrd0", the default, "nrd", "ucv", "bcv", "sj", "sj-ste", "sj-dpi").

...

Further arguments to the plot() call that opens each frame.

main, col, pal

The title, colours overriding the palette, and a blpal() name.

y

Numeric vector, one per value of x (plothex()).

bins

Whole number, default 30: hexagons across the range of x.

Value

The data drawn, invisibly: for plotrdg() and plotvln() the densities (grp, x, y), for plotvln() with the per-group median, q1 and q3 as an attribute summary, for plotecdf() the steps (grp, x, y), for plothex() the hexagon centres (x, y) and count, for plotgain() the curves (grp, population, value) with the Gini coefficients as an attribute gini.

Examples

set.seed(3)
x <- c(stats::rnorm(200), stats::rnorm(200, 2, 0.6))
g <- rep(c("before", "after"), each = 200)
plotrdg(x, g)

plotecdf(x, g)

plotvln(x, g)

plothex(x, x + stats::rnorm(400), bins = 20)

plotgain(stats::rexp(400), g)

try(plotrdg("a"))
#> Error : plotrdg: `x` must be a numeric vector