An example of a useful panel function for huge datasets
1 2 3 4  ipanel.smooth(x, y = NULL, pixs = 1, zmax = NULL,
ztransf = function(x) {x},
colramp = IDPcolorRamp, col = "black", lwd = 2,
span = 2/3, iter = 3, ...)

x,y 
Numeric vectors of the same length. 
pixs 
Size of pixel in x and ydirection in [mm] on the plotting device. When x and y are numeric, pixels are square. When x and y are factors, pixels are no longer square. The pixels are enlarged in xdirection. 
zmax 
Maximum number of counts per Pixel in the plot. When NULL, the density in the scatter plot is encoded from 0 pixel to maximum number of counts observed. zmax must be equal or larger than maximum number of counts found. 
ztransf 
Function to transform the number of counts per pixel,
which will be mapped by the function in 
colramp 
Color ramp to encode the number of counts within a pixel by color. 
col,lwd 
Color and line width of the “smoothed curve”. 
span 
the smoother span. This gives the proportion of points in the plot which influence the smooth at each value. Larger values give more smoothness. 
iter 
The number of robustifying iterations which should be performed. Using smaller values of iter will make lowess run faster. 
... 
Other graphical parameters as arguments to the

Rene Locher
lowess
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  r.lm < lm(Sepal.Length~Sepal.Width+Petal.Length+Petal.Width+Species,
data=iris)
par(mfrow=c(2,2),pty="s")
itermplot(r.lm, se = TRUE, partial.res=TRUE, smooth=ipanel.smooth,
lwd.smth = 3, pixs = 2, ask=FALSE)
if (require(SwissAir)) {
par(mfrow=c(1,1))
dat < AirQual[,c("ad.O3","ad.NOx","ad.T","ad.Td","ad.WS")]
dat$ad.O3 < log(dat$ad.O3)
dat$ad.NOx < log(dat$ad.NOx)
dat$ad.WS < log(dat$ad.WS)
pairs(dat,
panel=ipanel.smooth)
} else print("Package SwissAir is not available")

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.
All documentation is copyright its authors; we didn't write any of that.