R help - Jun 2004 - Covered Labels

Dear All!

 

How can I cope with overlapping or covered labels (covered by labels
from other data points) in plots?

 

 

 

Martina Renninger

 

 


	[[alternative HTML version deleted]]

On Wed, 2004-06-23 at 09:06, Martina Renninger wrote:
> Dear All!
> How can I cope with overlapping or covered labels (covered by labels
> from other data points) in plots?

Presuming that you are using text() to identify points in a plot, you
can use the 'cex' argument (which defaults to 1) to reduce the size of
the font. So in this case, try values <1, for example:

text(x, y, labels = YourText, cex = 0.8)

Possibly depending upon how many points you have, you can also adjust
the position of the label with respect to the data points by using
'adj', 'pos' and 'offset'.

See ?text for more information.

HTH,

Marc Schwartz

Martina Renninger wrote:
> Dear All!
>
>
>
> How can I cope with overlapping or covered labels (covered by labels
> from other data points) in plots?
This doesn't solve every such problem, but it has helped me in the past.
You will probably have to expand the xlim and ylim a bit to fit the extreme 
labels onto the plot.

Also, there seems to be a typo in the HTML help for dist(), mentioning 
as.matrix.dist() which should be as.matrix().

Jim
-------------- next part --------------
# thigmophobe returns the direction (as n|e|s|w) _away_ from
# the nearest point where x and y are vectors of 2D coordinates

thigmophobe<-function(x,y) {
 # get the current upper and lower limits of the plot
 plot.span<-par("usr")
 x.span<-plot.span[2] - plot.span[1]
 y.span<-plot.span[4] - plot.span[3]
 # if either axis is logarithmic, transform the values into logarithms
 if(par("xlog")) x<-log(x)
 if(par("ylog")) y<-log(y)
 # scale the values to the plot span
 # this avoids the numerically larger
 # axis dominating the distance measure
 x<-x/x.span
 y<-y/y.span
 # get the distance matrix as a full matrix
 xy.dist<-as.matrix(dist(cbind(x,y)))
 lenx<-length(x)
 nearest.index<-rep(0,lenx)
 for(index in 1:lenx)
  nearest.index[index]<-as.numeric(names(which.min(xy.dist[-index,index])))
 # get the x and y differences for each point to the nearest point
 xdiff<-x - x[nearest.index]
 ydiff<-y - y[nearest.index]
 # first set the east/west direction
 dir.ew<-ifelse(xdiff > 0,"e","w")
 # now do the north/south
 dir.ns<-ifelse(ydiff > 0,"n","s")
 dir.away<-ifelse(abs(xdiff)>abs(ydiff),dir.ew,dir.ns)
 return(dir.away)
}

# thigmophobe.labels positions labels at points so that they
# are most distant from the nearest other point, where the
# points are described as x and y coordinates.

thigmophobe.labels<-function(x,y,labels,...) {
 if(!missing(x) && !missing(y)) {
  lenx<-length(x)
  if(missing(labels)) labels<-as.character(1:lenx)
  dir.away<-thigmophobe(x,y)
  xadjust<-strwidth("M")
  yadjust<-strheight("M")*1.5
  text.adj<-rep(0,lenx)
  for(i in 1:lenx) {
   if(dir.away[i] == "n") {
    y[i]<-y[i]+yadjust
    text.adj[i]<-0.5
   }
   if(dir.away[i] == "e") {
    x[i]<-x[i]+xadjust
    text.adj[i]<-0
   }
   if(dir.away[i] == "s") {
    y[i]<-y[i]-yadjust
    text.adj[i]<-0.5
   }
   if(dir.away[i] == "w") {
    x[i]<-x[i]-xadjust
    text.adj[i]<-1
   }
  }
  text(x,y,labels,...)
 }
 else
  cat("Usage: thigmophobe.labels(x,y,labels=1:length(x),...)\n")
}