[igraph] Calculating Q-Spa modularity for spatially embedded networks

[T.O. Richardson]

Dear List,
I am relatively new to network analysis, and the networks I have are not 
simple: weighted, dynamic, undirected, and spatially-embedded.

According to this 2010 paper <http://arxiv.org/abs/1012.3409>, the classic 
Newman-Girvan modularity provides misleading community detection 
classifications when applied to spatially-embedded networks, because focus 
constraint and homophily are confounded in space.

So I was wondering (hoping) how easy it would be to implement the Q-spar 
modularity measure (equations5-6 in that paper) in R generally (e.g. using 
igraph or tnet).

I am no mathematician, but can code OK in R. From the look of equations, I 
am guessing that practically it will mean creating another matrix 
containing the pairwise distances between nodes, then.. well that's when it 
gets fuzzy.

Anyway, all pointers appreciated,

Best
Tom Richardson



-----------------------------------------------
Dept. of Maths and Stats, UWE and Ant Lab, Bristol

Room B79.
SoBS, Woodland Rd. BS8 1UG.
+44 (0)117 928 8443
----------------------------------------------------