TeachingDemos (version 2.12)

tkBrush: Change the Color and Styles of points interactively

Description

Creates a Tk window with a scatterplot matrix, then allows you to "brush" the points to change their color and/or style.

Usage

tkBrush(mat,hscale=1.75,vscale=1.75,wait=TRUE,...)

Value

Either NULL (if Wait=FALSE) or a list with components col and

pch corresponding to the state of the points.

Arguments

mat

A matrix of the data to plot, columns are variables, rows are observations, same as pairs

hscale

Passed to tkrplot

vscale

Passed to tkrplot

wait

Should the function wait for you to finish, see below

...

Additional arguments passed to the panel functions

Author

Greg Snow

Details

This function creates a Tk window with a pairs plot of mat, then allows you to interactively move a rectangle (the brush) over the points to change their color and plotting character.

The arrow keys can be used to change the size and shape of the brush. The left arrow makes the rectangle wider, the right makes it narrower. The up arrow key makes it taller, the right makes it shorter.

When the mouse button is not pressed the points inside the brush will change while in the brush, but return to their previous state when the brush moves off them. If the mouse button is pressed then the points inside the brush will be changed and the change will remain until a different set of conditions is brushed on them.

The style of the brushed points is determined by the values of the 2 entry boxes on the right side of the plot. You can specify the plotting character in the pch box, this can be anything that you would regularly pass to the pch argument of points, e.g. an integer or single character. You can specify the color of the brushed points using the color entry box, specify the name of any color recognized by R (see colors), if this box does not contain a legal color name then black will be used.

If wait is FALSE then the Tk window will exist independently of R and you can continue to do other things in the R window, in this case the function returns NULL. If wait is TRUE then R waits for you to close the Tk window (using the quit button) then returns a list with the colors and plotting characters resulting from your brushing, this information can be used to recreate the plot using pairs on a new graphics device (for printing or saving).

See Also

pairs,colors,points, the iplots package

Examples

if(interactive()){

# Iris dataset

out1 <- tkBrush(iris)

#  Now brush the points

pairs(iris, col=out1$col, pch=out1$pch)

# or

colhist <- function(x,...){
    tmp <- hist(x,plot=F)
    br <- tmp$breaks
    w <- as.numeric(cut(x,br,include.lowest=TRUE))
    sy <- unlist(lapply(tmp$counts,function(x)seq(length=x)))
    my <- max(sy)
    sy <- sy/my
    my <- 1/my
    sy <- sy[order(order(x))]
    tmp.usr <- par('usr'); on.exit(par(usr=tmp.usr))
    par(usr=c(tmp.usr[1:2],0,1.5))
    rect(br[w], sy-my, br[w+1], sy, 
       col=out1$col, # note out1$col is hardcoded here.
       border=NA)
    rect(br[-length(br)], 0, br[-1], tmp$counts*my)
}
pairs(iris, col=out1$col, pch=out1$pch, diag.panel=colhist)


# some spheres

s1 <- matrix(nrow=0,ncol=3)

while( nrow(s1) < 1000 ){
	tmp <- rnorm(3)
	if( sum(tmp^2) <= 1 ){
		s1 <- rbind(s1,tmp)
	}
}

s2 <- matrix(rnorm(3000), ncol=3)
s2 <- s2/apply(s2,1,function(x) sqrt(sum(x^2)))

tkBrush(s1, wait=FALSE)
tkBrush(s2, wait=FALSE)

# now paint values where var 2 is close to 0 in both plots 
# and compare the var 1 and var 3 relationship

}

Run the code above in your browser using DataCamp Workspace