[R] 3D plot not working as desired!

Duncan Murdoch murdoch at stats.uwo.ca
Tue Mar 8 09:36:07 CET 2005

On Tue, 8 Mar 2005 01:50:04 -0600, <kim83 at uiuc.edu> wrote :

>Hello R-users!
>I am trying to plot 3 vectors (x,y,z) of observations
>generated by mvrnorm in library(MASS).
>I tried plot3d in library(djmrgl) and scatterplot3d.
>But these program gives x,y,z axis which do not 
>intersect at the origin (0,0,0). 

The default in R is to create axes outside the range of your data, and
that's what djmrgl and scatterplot3d do.  In djmrgl you can control
the placement of the axes if you really want them in the middle of
your data, but why would you want that?? 

See the example in ?axis3d for details.
>I searched through all the graphics related packages
>for R like xgobi, ggobi, grid and more. Quite
>confusing which one to choose.
>Would anyone recommend any way to plot the 3-dimensional
>randomly generated numbers? It would be very nice if it
>can create 95% confidence region ellisoid around the
>observations, just like the 95% confidence ellipse around
>the bivariate normal random numbers.

I don't know a package that does that, but there probably is one.

Duncan Murdoch

More information about the R-help mailing list