Re: [igraph] degree distribution for bipartite network

[Tamas Nepusz]

> deg<-degree(g)[V(g)$type==FALSE]
> dd<-as.numeric(table(deg))/sum(as.numeric(table(deg)))
>
> Is it possible in python to find the degree distribution conditioned on a 
> node's attribute value?
Some of the graph methods that allow a vertex set as the first argument
can also be called like this:

g.vs.select(type=0).degree()

This simply translates to g.degree(g.vs.select(type=0)) and gives you
the degrees of vertices with type zero. You can then do whatever you
want with this list; e.g., feed them to a Histogram object:

degs = g.vs.select(type=0).degree()
hist = Histogram(bin_width=1)
hist << degs

Or, you can simply count the degrees in a dict:

hist = defaultdict(int)
for degree in g.vs.select(type=0).degree():
    hist[degree] += 1

Python does not have built-in classes for histograms/distributions,
unlike R, so you have to do this manually or create a helper function.
Matplotlib can then be used for plotting.

-- 
T.