|
| Mar 31 |
|
| Mar 31 |
|
| Mar 31 |
|
| Mar 31 |
|
| Mar 31 |
|
| Mar 31 |
|
| Mar 29 |
|
| Mar 28 |
|
| Mar 26 |
|
| Mar 26 |
|
Windows:
The easiest way to get started with GGobi is on Windows. First, you need to install R. See
http://groups.google.com/group/pdx-visualization/web/getting-started-with-the-r-programming-language
for the instructions. I'm going to assume you're not behind a firewall. If you are, you will need to ask your IT / firewall administrators for help and permission to do this.
Start up R in Windows. If you're on Vista, you'll probably want to use the "Run as Administrator" option. Then type at the ">" prompt:
source("http://www.ggobi.org/downloads/install.r")
It will ask you if you want to install "GTK+" and "GGobi". Answer "yes" to both. That's all you need to do.
Linux:
On Ubuntu or Debian, you should be able to do "apt-get install ggobi" and get it installed. On Fedora, you should be able to type "yum install ggobi" and get it installed. If you have either of these, please try it and let me know if it doesn't work. When you're done, you should have the latest version, 2.1.8, installed.
On openSUSE, the situation is a little more complicated. There is a pre-built package available in the openSUSE repositories, but it's a slightly older version (2.1.7) with a very annoying bug having to do with saving GGobi results in a format suitable for presentation graphics. As a result, I build GGobi from source on my openSUSE machines, and I recommend you do the same. If anybody else has openSUSE, let me know and I'll send you the build script I use. The instructions for building from source are at
http://ggobi.org/downloads/#linux
if you want to try it without help. You'll need to install compilers, "make", etc., and header files for the libraries that GGobi uses.
MacOS X:
Again, I have never done this, so please try it and let me know whether it works or not.
http://ggobi.org/downloads/#mac
Testing on a Social Network dataset:
Assuming you do get GGobi installed, start it up. You should get a control panel. Go into the "File" menu and open the "snetwork.xml" data file that comes with the package. The "open" menu should browse to the right directory. If it doesn't, let me know and I'll help you figure out where it is. Now go into the "Tools" menu. At the bottom, you should see an entry called "ggvis (MDS)". Go over to the "Run" tab and check the "Run MDS" box. A window should open up showing the "social network". Now in the "Edges" menu select "attach edge set (contacts)". It should "connect the dots". Go back to the main GGobi control panel and select "Move Points" in the "Interaction" menu. Now you should be able to mouse over a node and see the person's name! Drag a point somewhere and the MDS code will move it back where you found it! Uncheck the "Run MDS" box and it should let you drag the nodes around without restoring them to their computed positions.
There's another, probably better, version of this dataset at
http://ggobi.org/book/data/snetwork.xml
Here's a brief paper on the subject and a brief slide presentation: http://www.ci.tuwien.ac.at/Conferences/DSC-2003/Proceedings/SwayneEtAl.pdf If you want to import one of your own social networks into XML, the main XML format documentation is at
The actual manual is at
http://ggobi.org/docs/manual.pdf
|
| ||||||||||||||||||||||||||||
| Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy |
| ©2010 Google |