igraph-help
[Top][All Lists]

## [igraph] question about itereted operations and egonetworks

 From: Umberto77 Subject: [igraph] question about itereted operations and egonetworks Date: Tue, 16 Apr 2013 17:36:51 +0200

Hi all,
Hi Tamas and Gabor,

I confess I were using Ucinte for SNA until I moved to igraph in R. The reason why I did it, although I'm not used to R and I find it quite puzzling given my background (I'm an agronomist working in rural development issues, try to imagine my efforts) is that I was looking for a tool that could iterate procedures normally time consuming in Ucinet.

For example, I'd like to calculate a set of egonetwork measures performed by Ucinet (http://faculty.ucr.edu/~hanneman/nettext/C9_Ego_networks.html) over egonetworks extracted from a time series of the same network (10 years = 10 networks).
The measures are: size, ties, pairs, density,Average distance, diameter, number of weak components, 2steps reachability, reach efficiency, broakerage, normalized brokerage, egobetweenness, normalized egobetweeness.

Let's focus on density.

I figure it out I can use "graph.neighborhood" to creat a list of egonetworks for a given network. Than, I can use "lappy" to iterate "graph.density" over each egonetwork of my previous list.

The problem is that using graph.neighborhood I get an output like this:

[[65]]
IGRAPH DN-- 6 15 --
+ attr: name (v/c), value (e/n)

[[66]]
IGRAPH DN-- 3 2 --
+ attr: name (v/c), value (e/n)

where egos in my list are identified by a number in squared brackets. Since I need to work on a time series, I need to keep trace of the name of the egos, so that to build an egonetwork density time series for each of my nodes across 10 years.

As for "graph.density" it calculates density as ratio between number of ties and all the possible ties in the (ego)network, while I need to calculate it as number of ties over number of pairs. I was wondering if there's a direct function to perform it in place of "graph.density".
I'll keep working to fin out how to calculate the rest of the (Ucinet) egonetwork measures listed above, but I've still one last question.

What if I create a list of networks, including all my 10 years, and than use "lapply" for "graph.neighborhood" and than for "graph.density"?

I can see my questions relates to R usage more than Igraph, I apologize.

Thanks,

Umberto

----
Caselle da 1GB, trasmetti allegati fino a 3GB e in piu' IMAP, POP3 e SMTP autenticato? GRATIS solo con Email.it