DescTools (version 0.99.19)

PlotViolin: Plot Violins Instead of Boxplots

Description

This function serves the same utility as side-by-side boxplots, only it provides more detail about the different distribution. It plots violins instead of boxplots. That is, instead of a box, it uses the density function to plot the density. For skewed distributions, the results look like "violins". Hence the name.

Usage

PlotViolin(x, ...)
"PlotViolin"(x, ..., horizontal = FALSE, bw = "SJ", na.rm = FALSE,  names = NULL, args.boxplot = NULL) 
"PlotViolin"(formula, data, subset, na.action, ...)

Arguments

x

Either a sequence of variable names, or a data frame, or a model formula

horizontal

logical indicating if the densityplots should be horizontal; default FALSE means vertical arrangement.

bw

the smoothing bandwidth (method) being used by density. bw can also be a character string giving a rule to choose the bandwidth. See bw.nrd. The default, has been switched from "nrd0" to "SJ", following the general recommendation in Venables & Ripley (2002). In case of a method, the average computed bandwidth is used.

na.rm

logical, should NAs be omitted? The density-function can't do with missings. Defaults to FALSE.

names

a vector of names for the groups.

formula

a formula, such as y ~ grp, where y is a numeric vector of data values to be split into groups according to the grouping variable grp (usually a factor).

data

a data.frame (or list) from which the variables in formula should be taken.

subset

an optional vector specifying a subset of observations to be used for plotting.

na.action

a function which indicates what should happen when the data contain NAs. Defaults to getOption("na.action").

...

The dots are passed to polygon. Notably, you can set the color to red with col="red", and a border color with border="blue"

args.boxplot

list of arguments for a boxplot to be superposed to the densityplot. By default (NULL) a black boxplot will be drawn. Set this to NA to suppress the boxplot.

Value

If a boxplot was drawn then the function returns a list with the following components:

References

The code is based on the boxplot function from R/base.

See Also

boxplot, PlotMultiDens , density

Examples

# make a "violin"
x <- c(rnorm(100), rnorm(50,5))

PlotViolin(x, col = "brown")

par(mfrow=c(1,2))
f <- factor(rep(1:5, 30))
# make a quintet. Note also choice of bandwidth
PlotViolin(x ~ f, col = SetAlpha("steelblue",0.3), bw = "SJ", main="Vertical")

# and the same, but in horizontal arrangement
PlotViolin(x ~ f, col = SetAlpha("steelblue",0.3), bw = "SJ", horizontal = TRUE, 
  las=1, main="Horizontal")


# example taken from boxplot
boxplot(count ~ spray, data = InsectSprays, col = "lightgray", main="Boxplot")
PlotViolin(count ~ spray, data = InsectSprays, col = "lightgray", main="Violinplot")


# groupwise densityplots defined the same way as in boxplot
boxplot(len ~ supp*dose, data = ToothGrowth,
        main = "Guinea Pigs' Tooth Growth",
        xlab = "Vitamin C dose mg", ylab = "tooth length",
        col=c("yellow", "orange"), lty=c(1,2)
)

b <- PlotViolin(len ~ supp*dose, data = ToothGrowth,
           main = "Guinea Pigs' Tooth Growth",
           xlab = "Vitamin C dose mg", ylab = "tooth length",
           col=c("yellow", "orange"), lty=c(1,2)
)
# use points, if the medians deserve special attention
points(x=1:6, y=b$stats[3,], pch=21, bg="white", col="black", cex=1.2)