Graph Database for Collaborative Communities - Community-Built Databases: Research and Development - page 221

9.4 Social Network Extraction from Relational Database
Using a Graph Database
A Social Network is an explicit representation of relationships between people,
groups, organizations, computers, or other entities and it is modeled by a graph
(see Sect. 9.2.1.3). There are many ways to obtain a social network. The app-
roaches presented in the literature on social network extraction use a specific type
of data source to extract people and relations among them [ 36 ]. Most of these data
sources come from the Web. However, some problems related to the extraction of
a social network from various information sources available on the World Wide
Web still remain. First, a general problem is the identification of people because
of different naming standards or the same names assigned to different persons.
The social context and the type of social interactions among people within these
information sources need to be carefully analyzed in order to obtain a meaningful
understanding of the underlying Social Network structure. Moreover, data from
the Web are often not very reliable because anyone can add information; also, in
some cases, we cannot easily collect information from the Web due to privacy
issues.
Nevertheless, in the context of business, important expertise information about
people is not stored on the Web. Such information is stored in files, databases, and
especially relational databases. A relational database is a rich source of data, but it
is not appropriate for storing and manipulating social network data. Indeed, the
relational model was intended for simple record-type data with a structure known in
advance. The schema is fixed and extensibility is difficult. Thus, it might require
very sophisticated and expensive operations, such as renormalization and reindex-
ing, which might not be performed automatically. Schema renormalization in such
cases is neither desirable nor easy to perform. The standard query and transforma-
tion language for the relational database is SQL, which does not support paths,
neighborhoods, and queries that address connectivity (an exception is transitivity).
These graph features will facilitate the application of a social network analysis
algorithm. Also, it will allow response to queries such as who owns the information,
who has the leadership, and who is an expert in a particular domain. Such informa-
tion is very important for business applications. Then, enterprises need to extract
their social network from the existing relational database to store, update, and
retrieve information in a simple way such as graphs. On the other hand, extracting
a social network from a relational database is not just a translation of a relational
database into a simple graph structure. The resulting social network should contain
detailed information about people and their relations. As we have shown in the
previous section, a graph database can be a good representation for social networks
and facilitate its querying. There are many approaches that transform relational
databases to other structures having graph-like features such as RDF, XML, or even
ontology, but not into a graph database. Hence, in this section, we will present our
approach to transforming a relational database to a social network using a graph
database.