Information Technology Reference
In-Depth Information
the second case of social media, it is people. There are many other manifestations of
interconnected structures in the world where connection patterns can be described
with networks.
Examples of networks that exist in the real world include the world wide web [2],
transport networks [9], cell biology networks [5], software networks [26] and lin-
guistic networks [52]. Albert and Barabasi provide further examples of networks
and their properties in [1].
By contrast to inanimate networks, social networks define a subset of networks
that involve human interaction and human actors. In social networking terms, a net-
work is defined as a set of actors (objects or
nodes
in inanimate networks) and
a set of ties (connections or
edges
) between those actors [51]. Typical examples
of actors are persons, organisations or countries. Ties represent relationships be-
tween the actors. Often relationships link actors based on some kind of interac-
tion, which explains why links are described by verbs including; married-to, son-of,
sought-advice-from, plays-with, emailed-to, donated-money-to, participated-with
and talks-to.
1.1
Implicit and Explicit Network Data Sources
The distinguishing feature of network data is that this data is
relational
as opposed
to simply
attribute
based. Relational data is based upon connections between data
elements, whereas attribute data is based on the properties of the data elements and
each data element is independent. The analytical techniques used for the two types
of data also vary. Relational data is usually analysed using Social Network Analysis
(SNA) [55], while attribute data is analysed using variable analysis.
Relational data can be either
explicitly
collected or
implicitly
extracted from the
raw data. Explicit network data sources unambiguously define both the actors and
the relationships between those actors. These data sources are typically collected
with the explicit intent of analysing the data using social network analysis tech-
niques instead of variable based analysis. Most manual data collection methods in
sociological studies collect data using self reporting surveys where the relationships
and the nature of the relationship is specified clearly in the survey [39, 59, 17].
With implicit data sources the actors, or more often their relationships, are not
explicitly defined. These data sets are not collected with the intention of studying
the relationships between actors, but intended to be analysed independently using
variable analysis. The Al Qaeda network [40] extracted by Krebs, is a clear example
of an implicit network. This network was constructed from media reports follow-
ing the attacks of September 11 and was generated from a number of data sources,
including news articles were the relationships are implied, or sometimes even spec-
ulative. In theory, social networking web-sites [16] form explicit networks as the
focus of the data is on the relationships. The relationships between friends are ex-
plicitly defined and subsequently confirmed by both actors. In reality however, the
abused definition of the friendship relationship can limit the sociological signifi-
cance of the underlying social network. Familial, social, workplace and even sexual
