boxplot2D {aplpack} | R Documentation |
boxplot2D computes summary statistics of a one dimensional projection of a two dimensional data set and plots a sloped boxplot of the statistics into the scatterplot of the two dimensional data set.
boxplot2D(xy, add.to.plot = TRUE, box.size = 10, box.shift = 0, angle = 0, angle.type = "0", tukey.style = TRUE, coef.out = 1.5, coef.h.out = 3, design = "sl", na.rm=FALSE, ...)
xy |
(nx2) -matrix, two dimensional data set |
add.to.plot |
if TRUE the boxplot is added to
the actual plot of the graphics device |
box.size |
height of the box (of the boxplot) |
box.shift |
shift of boxplot perpendicular to the projection direction |
angle |
direction of projection in units defined by angle.type |
angle.type |
"0" : angle in (0,2*pi),
"1" : clock-like: angle.typ.0==2*pi*angle.typ.1/12,
"2" : degrees: angle.typ.0==2*pi*angle.typ.2/360,
"3" : by fraction: delta.y/delta.x
|
tukey.style |
if TRUE outliers are defined as described
in Tukey (1977)
|
coef.out |
outliers are values that are more than
coef.out*boxwidth away from the box, default:
coef.out=1.5 |
coef.h.out |
heavy outliers are values that are more
than coef.h.out*boxwidth away from the box,
default: coef.h.out=3 |
design |
if sl then parallelogram else box |
na.rm |
if TRUE 'NA' values are removed otherwise exchanged by mean |
... |
additional graphical parameters |
version 08/2003
Peter Wolf
Tukey, J. Exploratory Data Analysis. Addison-Wesley, 1977.
xy<-cbind(1:100, (1:100)+rnorm(100,,5)) plot(xy,xlim=c(-50,150),ylim=c(-50,150)) boxplot2D(xy,box.shift=-30,angle=3,angle.typ=1) boxplot2D(xy,box.shift=20,angle=1,angle.typ=1) boxplot2D(xy,box.shift=50,angle=5,angle.typ=1)