NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.

To learn about what's going on, see About the Archive Site.

Re: Shapes in space

Hi David,

> I have a small VisAD program that will plot particles in space like
> this:
> 
> (index) --> ( x, y, z, value )
> 
> I run my data files through a cluster algrothm and have textual
> information concerning the clusters, but I would like to emphasize these
> clusters in the visualization. I think I would like to draw boundaries
> around these cluster and was wondering if there is a method in VisAD
> that will allow me to accomplish this?

VisAD allows you to define closed surfaces (for a 3-D
boundary) as an Irregular3DSet with manifold dimension
= 2, but does not provide any method for computing that
boundary. Your own cluster analysis code would need to
do that.

> P.S. any other suggestions on how I could visualize my clustering
> results would be appreciated.

You might assign an integer cluster_index to the points in
each cluster, then produce data with MathType:

  (index -> (x, y, z, value, cluster_index))

You can then display with a ScalarMap cluster_index -> RGB
so the points in each color have a distinct color. You may
also use a ScalarMap cluster_index -> SelectRange so that
you can interactively select the points in a single cluster
or a range of cluster_index values. If you reorganize your
data as:

  (cluster_index -> (index -> (x, y, z, value)))

Then you can display with a ScalarMap cluster_index -> SelectValue
which will allow you to select single clusters with very fast
response.

Good luck,
Bill
---------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hibbard@xxxxxxxxxxxxxxxxx  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html


  • 2002 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the visad archives: