compositions (version 2.0-2)

boxplot: Displaying compositions and amounts with box-plots

Description

For the different interpretations of amounts or compositional data, a different type of boxplot is feasible. Thus different boxplots are drawn.

Usage

# S3 method for acomp
boxplot(x,fak=NULL,…,
                         xlim=NULL,ylim=NULL,log=TRUE,
                         panel=vp.logboxplot,dots=!boxes,boxes=TRUE,
                          notch=FALSE,
                          plotMissings=TRUE,
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                missingInfo,c("NM","TM",cn))
                          )
# S3 method for rcomp
boxplot(x,fak=NULL,…,
                         xlim=NULL,ylim=NULL,log=FALSE,
                         panel=vp.boxplot,dots=!boxes,boxes=TRUE,
                          notch=FALSE,
                          plotMissings=TRUE,
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                missingInfo,c("NM","TM",cn)))
# S3 method for aplus
boxplot(x,fak=NULL,…,log=TRUE,
                          plotMissings=TRUE,
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                   missingInfo,
                                                   names(missingInfo)))
# S3 method for rplus
boxplot(x,fak=NULL,…,ylim=NULL,log=FALSE,
                          plotMissings=TRUE,
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                   missingInfo,
                                                   names(missingInfo)))
vp.boxplot(x,y,...,dots=FALSE,boxes=TRUE,xlim=NULL,ylim=NULL,log=FALSE,
                          notch=FALSE,plotMissings=TRUE,
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                   missingInfo,c("NM","TM",cn)),
                          missingness=attr(y,"missingness") ) 
vp.logboxplot(x,y,...,dots=FALSE,boxes=TRUE,xlim,ylim,log=TRUE,notch=FALSE,
                          plotMissings=TRUE, 
                          mp=~simpleMissingSubplot(missingPlotRect,
                                                   missingInfo,c("NM","TM",cn)),
                          missingness=attr(y,"missingness"))

Arguments

x

a data set

fak

a factor to split the data set, not yet implemented in aplus and rplus

xlim

x-limits of the plot.

ylim

y-limits of the plot.

log

logical indicating whether ploting should be done on log scale

panel

the panel function to be used or a list of multiple panel functions

…

further graphical parameters

dots

a logical indicating whether the points should be drawn

boxes

a logical indicating whether the boxes should be drawn

y

used by pairs

notch

logical, should the boxes be notched?

plotMissings

Logical indicating that missings should be displayed.

mp

A formula providing a function call, which will be evaluated within each panel with missings to plot the missingness situation. The call can use the variables missingPlotRect, which provides a rectangle to plot the information to in a par("usr") like specification. In the rX is the current data

missingness

The missingness information as a result from missingType of the full data information the panels could base there missing plots on.

Details

boxplot.aplus and boxplot.rplus are wrappers of bxp, which just take into account the possible logarithmic scale of the data.

boxplot.acomp and boxplot.rcomp generate a matrix of box-plots, where each cell represents the difference between the row and column variables. Such difference is respectively computed as a log-ratio and a rest.

vp.boxplot and vp.logboxplot are only used as panel functions. They should not be directly called.

See Also

plot.acomp, qqnorm.acomp

Examples

# NOT RUN {
data(SimulatedAmounts)
boxplot(acomp(sa.lognormals))
boxplot(rcomp(sa.lognormals))
boxplot(aplus(sa.lognormals))
boxplot(rplus(sa.lognormals))
# And now with missing!!!
boxplot(acomp(sa.tnormals))

# }