[R] 3d scatterplot with error bars

[hadley wickham]

> Is there anyway to produce a 3D scatterplot with error bars in the x,y,z
> directions?  I have searched around and know of scatterplot3d but did
> not see any way to put error bars on the points.  Any ideas?

It may be possible, but do you think that's the best solution for your
problem?  Given that any 3d plot must be projected into two
dimensions, how on earth will people be able to compare the relative
magnitude in each direction or between points?

One could imagine drawing error ellispoids (since you might as well
include covariance information as well) but I would imagine that the
tight connection between the view of the data and the two-D projection
of the ellipsoid (representing a particular linear combination of the
x, y and z errors) would make it very difficult to see what's going
on.  Not to mention the difficulties with near points obscuring far
points, etc etc.

Hadley

-- 
http://had.co.nz/