figure 1: the giant component in the GsJ network (data courtesy of marc sageman9).
the terrorists belong to one of four groups: al Qaeda or central staff (pink), core
arabs (yellow), maghreb arabs (blue), and southeast asians (green). each circle
represents one or more terrorist activities (such as the september 11 attacks and
the Bali bombing) as noted.
each other. It also has been shown that
the greater tendency for clustering in
metabolic networks corresponds to
the organization of functional modules in cells, contributing to the behavior and survival of organisms. In
addition, networks with scale-free
properties are highly robust against
random failure and errors but notably
vulnerable to targeted attacks. 5
Bali Bombing, 2002
9/11 Attacks, 2001
table 1: statistics we used for studying network topology.
the average of the lengths of the shortest paths
between all pairs of nodes in a network.
Average clustering coefficient, C11 the average of all individual clustering coefficients, Ci,
which is the number of links that actually exist among
node i’s neighbors over the possible number of links
among these neighbors.
Average degree, <k> 10 the average of all individual degrees, ki, which is
the number of links that node i has.
the probability that an arbitrary node has exactly k links.
the number of links that actually exist over the possible
number of links in a network.
Assortativity, r8 the Pearson correlation between the degrees of two
Global efficiency, e4 the average of the inverses of the lengths of the shortest
paths over all pairs of nodes in a network.
Average Path length, l1
degree distribution, p(k) 1
link density, d10
acterized by the power-law degree distribution, meaning that while a large
percentage of nodes in the network
has just a few links, a small percentage of the nodes have a large number
of links. 1 Scientists conjecture that
scale-free networks evolve following
the self-organizing principle, where
growth and preferential attachment
play a key role in the emergence of
the power-law distribution. Preferential attachment implies that the more
links a node has, the more new links
it is able to attract, manifesting the
Analyzing the topology of complex
systems has important implications
for our understanding of nature and
society. Research has shown that the
function of a complex system may be
affected to a great extent by its network
topology. 1 For instance, the Web’s
short average path length makes cyberspace a convenient, navigable system in which any two Web pages are
(on average) only 19 clicks away from
methods and Data
To understand the topology and function of dark networks we studied four
terrorist- and criminal-related networks:
Global Salafi Jihad (GSJ). 9 This terrorist network’s 366 members (see
Figure 1) include some from Osama
bin Laden’s Al Qaeda, connected by,
perhaps, kinship, friendship, religious ties, and relationships formed
after they joined. The GSJ data was
provided to us by Marc Sageman, a forensic psychiatrist in private practice
in Philadelphia and author of
Understanding Terror Networks. 9 The network was constructed entirely from
open-source data, including publicly
available documents and transcripts
of court proceedings and press, scholarly, and Web articles. Sageman scrutinized and cross-validated the information about all nodes (terrorists)
and links (relationships). However, as
he pointed out in his book, the data
is also subject to several limitations.
First, the members in the network
may not be a representative sample of
the global Salafi jihad. The data may
be biased toward leaders and members captured or identified in attacks.
Second, because most of the sources
were based on retrospective accounts,
the data may be subject to self-report-ed bias. Despite these limitations, the
data provides stunning insight into
clandestine terrorist organizations.
Meth World. In trafficking illegal
methamphetamines, 12 this network
consisted of 1,349 criminals traced
and investigated by the Tucson Police
Department from 1985 to 2002. Because no information about the social
relationships among them is directly
available, we were granted access to
the police databases and retrieved all
the crime incidents in which these
people were involved from 1985 to
2002. We created a link between any