Skip to content
Prev 75418 / 398502 Next

scatter plot

Here is a hack

 par(mfrow=c(1,2))

 x  <- rnorm(1000)
 y  <- x + rnorm(1000)
 xp <- (rank(x)-1)/(length(x)-1)

 plot( x, y )
 plot( xp, y )

But do notice that by using percentiles, it spreads the 'x' values
evenly. This may be important for points at the extremes.

Regards, Adai
On Thu, 2005-08-11 at 15:40 -0700, Dean Sonneborn wrote: