What Social Network Analysis software do you use?
See a the poll here by Gabriel Rossman at Code and Culture.
I voted for R and ‘igraph’.
If you use R you are getting access to all the other wonderful things that come with R. Using specialized package, like Pajek, UCINET etc requires constant going back and forth between network software and some other general analytical software (sas, spss, what have you).
I also use both R packages. Mainly ‘igraph’ (and that is what I voted for) because I find the network data manipulation much easier than in ‘network’. I believe it is also much more efficient computationally than ‘network’ (Gabor Csardi showed couple of comparisons on UseR 2009). On the other hand with ‘network’ you get access to ERGM and all the other statnet packages written by people at Stats Dept at University of Washington. So igraph/network is not an easy choice.
I’m actually working on reliable routines that will convert igraph objects to network and vice versa. Perhaps an ideal situation would be to have a an umbrella interface to igraph, network, graph etc. very similar to DBI as an umbrella over RMySQL, RSQLite etc…..
Brokering the Closure 
Pretty please with sugar on top!
You mean the syntactic sugar? 😉
Michael, it sounds like I use a fairly similar set of packages to you.
I’m also planning on checking out the RSiena package shortly.
I wrote a post on R and SNA a while back, that might interest some people:
http://jeromyanglim.blogspot.com/2009/08/social-network-analysis-resources-for-r.html
Others in my department have also developed useful friendly stand alone social network software for ERGM and such:
http://www.sna.unimelb.edu.au
Ah yes, the PNet package. I forgot about it. I have to admit I have never used it. Before ‘statnet’ and ‘ergm’ was around I used StocNet by Tom Snijders to fit p-star models. How does it compare to statnet in terms of modeling capabilities? (apart from the obvious differences that pnet is specialized and have a java gui)
Thanks for the link to your post. I’m actually subscribing your blog through Google Reader, so I now that post already 😉
Hi guys,
I have two remotely related questions:
1. So far, I have done SNA using the network package. Now I got strange results (different results each time I ran the analysis) for brokerage scores (Gould and Fernandez 1989). Is there a way to get brokerage scores in igraph?
2. I want to generate some random networks with specific vertex attributes – any suggestions?
Thanks!
Philipp
Hey Philipp,
It is very strange that you are getting different results each time you calculate brokerage scores. It’s surprising because the procedure is completely deterministic so it should return exactly the same result every time. Are you sure you didn’t make some other kind of mistake?
As for your second question: I’m not sure if I understood correctly, but you want to generate some random networks in which tie probability is in some way related to a vertex attribute(s)? In that case I would probably do this via sampling from some Exponential Random Graph Model (package ‘ergm’) which includes vertex-attribute-related effects… Depends on exactly what kind of networks you would like to generate (?)
Michal,
many thanks!
You were right regarding the brokerage. I simply assigned a wrong class membership vector – it works now.
Regarding my second question:
True, I want to generate graphs of vertices with a male – female attribute (though, I would assign the same probability for both). But I guess I now know where to look.
Again, thanks!
Sure, np 😉
AFAIK there is no function to calculate brokerage scores for igraph objects. It would be useful though. I put it on my TOCODE list.