***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
since it looks like email networks is becoming a hot topic in the list
we would like to point you to a couple of PDF documents where we
analyze in detail the email network of a mid-sized university (composed of
1700 nodes including academic, administration staff and PhD students).
The main conclusions derived from these works are:
- The e-mail network is indeed a social network as pointed out by Prof.
Krackhardt that reveals the informal network behind the formal chart of any
- The community structure can be identified by the algorithm proposed by
Girvan and Newman (PNAS 99, 7821-7826, 2002). It determines perfect
communities (except ambiguous nodes) and provides a binary tree where the
leaders appear at the ending leaves.
- The community structure presents a stricking self-similar structure, that
it is not an artifact of the community identification algorithm as we show
article. It suggest that a self-organized universal mechanism is driving
the informal network formation.
The first document, more "physicist oriented", was posted in the Los Alamos
archive on 22 Nov 2002, and is currently being considered to be published in
a physics journal.
The second document, which has been written for a wider audience, exploits
the details that can be of interest from an organizational point of view:
identification of communities (schools, departments, research groups, ...)
in the university, leadership role, relation between the centers, distances
in the social network, and so on.
Both documents can be downloaded from our group of Physics of Complex
Systems web pages
This work was presented by A. Arenas in the 7th Granada Seminar
on Computational and Statistical Physics (held in Granada, Spain, 2-7
September 2002) http://ergodic.ugr.es/cp/pages/7th2002.htm
The recently paper written by Huberman et al. presents a nice algorithm
that corroborate our results in a company half of the size of our
university. Nevertheless, the algorithm of Girvan and Newman discriminates
communities up to the size of individual persons and is a matter of choice
to define the minimal size we assign to a community. We argue that the GN
algorithm complemented with the full visualization of the binary tree is an
excellent tool for management purposes.
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.