% File src/library/graphics/man/bxp.Rd % Part of the R package, http://www.R-project.org % Copyright 1995-2007 R Core Team % Distributed under GPL 2 or later \name{bxp} \alias{bxp} \title{Draw Box Plots from Summaries} \description{ \code{bxp} draws box plots based on the given summaries in \code{z}. It is usually called from within \code{\link{boxplot}}, but can be invoked directly. } \usage{ bxp(z, notch = FALSE, width = NULL, varwidth = FALSE, outline = TRUE, notch.frac = 0.5, log = "", border = par("fg"), pars = NULL, frame.plot = axes, horizontal = FALSE, add = FALSE, at = NULL, show.names = NULL, \dots) } \arguments{ \item{z}{a list containing data summaries to be used in constructing the plots. These are usually the result of a call to \code{\link{boxplot}}, but can be generated in any fashion.} \item{notch}{if \code{notch} is \code{TRUE}, a notch is drawn in each side of the boxes. If the notches of two plots do not overlap then the medians are significantly different at the 5 percent level.} \item{width}{a vector giving the relative widths of the boxes making up the plot.} \item{varwidth}{if \code{varwidth} is \code{TRUE}, the boxes are drawn with widths proportional to the square-roots of the number of observations in the groups.} \item{outline}{if \code{outline} is not true, the outliers are not drawn.} \item{notch.frac}{numeric in (0,1). When \code{notch = TRUE}, the fraction of the box width that the notches should use.} \item{border}{character or numeric (vector), the color of the box borders. Is recycled for multiple boxes. Is used as default for the \code{boxcol}, \code{medcol}, \code{whiskcol}, \code{staplecol}, and \code{outcol} options (see below). } \item{log}{character, indicating if any axis should be drawn in logarithmic scale, as in \code{\link{plot.default}}.} \item{frame.plot}{logical, indicating if a \sQuote{frame} (\code{\link{box}}) should be drawn; defaults to \code{TRUE}, unless \code{axes = FALSE} is specified.} \item{horizontal}{logical indicating if the boxplots should be horizontal; default \code{FALSE} means vertical boxes.} \item{add}{logical, if true \emph{add} boxplot to current plot.} \item{at}{numeric vector giving the locations where the boxplots should be drawn, particularly when \code{add = TRUE}; defaults to \code{1:n} where \code{n} is the number of boxes.} \item{show.names}{Set to \code{TRUE} or \code{FALSE} to override the defaults on whether an x-axis label is printed for each group.} \item{pars,\dots}{\link{graphical parameters} (etc) can be passed as arguments to this function, either as a list (\code{pars}) or normally(\code{\dots}), see the following. (Those in \code{\dots} take precedence over those in \code{pars}.) Currently, \code{yaxs} and \code{ylim} are used \sQuote{along the boxplot}, i.e., vertically, when \code{horizontal} is false, and \code{xlim} horizontally. \code{xaxt}, \code{yaxt}, \code{las}, \code{cex.axis}, and \code{col.axis} are passed to \code{\link{axis}}, and \code{main}, \code{cex.main}, \code{col.main}, \code{sub}, \code{cex.sub}, \code{col.sub}, \code{xlab}, \code{ylab}, \code{cex.lab}, and \code{col.lab} are passed to \code{\link{title}}. In addition, \code{axes} is accepted (see \code{\link{plot.window}}), with default \code{TRUE}. The following arguments (or \code{pars} components) allow further customization of the boxplot graphics. Their defaults are typically determined from the non-prefixed version (e.g., \code{boxlty} from \code{lty}), either from the specified argument or \code{pars} component or the corresponding \code{\link{par}} one. \describe{ \item{boxwex:}{a scale factor to be applied to all boxes. When there are only a few groups, the appearance of the plot can be improved by making the boxes narrower. The default depends on \code{at} and typically is \eqn{0.8}.} \item{staplewex, outwex:}{staple and outlier line width expansion, proportional to box width; both default to 0.5.} \item{boxlty, boxlwd, boxcol, boxfill:}{box outline type, width, color, and fill color (which currently defaults to \code{col} and will in future default to \code{par("bg")}).} \item{medlty, medlwd, medpch, medcex, medcol, medbg:}{median line type, line width, point character, point size expansion, color, and background color. The default \code{medpch = NA} suppresses the point, and \code{medlty = "blank"} does so for the line. Note that\code{medlwd} defaults to \eqn{3\times}{3x} the default \code{lwd}.} \item{whisklty, whisklwd, whiskcol:}{whisker line type (default: \code{"dashed"}), width, and color.} \item{staplelty, staplelwd, staplecol:}{staple (= end of whisker) line type, width, and color.} \item{outlty, outlwd, outpch, outcex, outcol, outbg:}{outlier line type, line width, point character, point size expansion, color, and background color. The default \code{outlty = "blank"} suppresses the lines and \code{outpch = NA} suppresses points.} } }%.../pars } \note{ When \code{add = FALSE}, \code{xlim} now defaults to \code{xlim = range(at, *) + c(-0.5, 0.5)}. It will usually be a good idea to specify \code{xlim} if the "x" axis has a log scale or \code{width} is far from uniform. } \value{ An invisible vector, actually identical to the \code{at} argument, with the coordinates ("x" if horizontal is false, "y" otherwise) of box centers, useful for adding to the plot. } \author{ The R Core development team and Arni Magnusson (then at U Washington) who has provided most changes for the box*, med*, whisk*, staple*, and out* arguments. } \references{ Becker, R. A., Chambers, J. M. and Wilks, A. R. (1988) \emph{The New S Language}. Wadsworth & Brooks/Cole. } \examples{ require(stats) set.seed(753) (bx.p <- boxplot(split(rt(100, 4), gl(5, 20)))) op <- par(mfrow = c(2, 2)) bxp(bx.p, xaxt = "n") bxp(bx.p, notch = TRUE, axes = FALSE, pch = 4, boxfill = 1:5) bxp(bx.p, notch = TRUE, boxfill = "lightblue", frame = FALSE, outl = FALSE, main = "bxp(*, frame= FALSE, outl= FALSE)") bxp(bx.p, notch = TRUE, boxfill = "lightblue", border = 2:6, ylim = c(-4,4), pch = 22, bg = "green", log = "x", main = "... log = 'x', ylim = *") par(op) op <- par(mfrow = c(1, 2)) ## single group -- no label boxplot (weight ~ group, data = PlantGrowth, subset = group == "ctrl") ## with label bx <- boxplot(weight ~ group, data = PlantGrowth, subset = group == "ctrl", plot = FALSE) bxp(bx, show.names=TRUE) par(op) %% examples for new (S+ like) features z <- split(rnorm(1000), rpois(1000, 2.2)) boxplot(z, whisklty = 3, main = "boxplot(z, whisklty = 3)") ## Colour support similar to plot.default: op <- par(mfrow = 1:2, bg = "light gray", fg = "midnight blue") boxplot(z, col.axis = "skyblue3", main = "boxplot(*, col.axis=..,main=..)") plot(z[[1]], col.axis = "skyblue3", main = "plot(*, col.axis=..,main=..)") mtext("par(bg=\"light gray\", fg=\"midnight blue\")", outer = TRUE, line = -1.2) par(op) ## Mimic S-Plus: splus <- list(boxwex = 0.4, staplewex = 1, outwex = 1, boxfill = "grey40", medlwd = 3, medcol = "white", whisklty = 3, outlty = 1, outpch = NA) boxplot(z, pars = splus) ## Recycled and "sweeping" parameters op <- par(mfrow = c(1,2)) boxplot(z, border = 1:5, lty = 3, medlty = 1, medlwd = 2.5) boxplot(z, boxfill = 1:3, pch = 1:5, lwd = 1.5, medcol = "white") par(op) ## too many possibilities boxplot(z, boxfill = "light gray", outpch = 21:25, outlty = 2, bg = "pink", lwd = 2, medcol = "dark blue", medcex = 2, medpch = 20) } \keyword{aplot}