Implementation of Social Network Analysis in Enterprise Architecture Modeling and Analysis (Pierre Gerardi)
Project description
Enterprise Architecture (EA) is a field that studies the structural and dynamic composition and properties of an organisation’s ‘system’ of goals, strategies, business organisation (roles, responsibilities, processes, functions, services, products, ..), information systems (data and applications), information technology, other infrastructure (facilities, machinery), and implementation and migration plans (programs, projects, work packages). Enterprise architects are responsible for the coherence of all these organisational assets and knowledge items and help to steer the organisation in the right direction, supporting its strategic initiatives with their knowledge, skills and toolset.
Social Network Analysis (SNA) is an increasingly popular technique in the social sciences as well as in other disciplines. Using SNA the network of interactions between actors can be modelled, measured and analysed. Numerous graph-based metrics have been defined for identifying, for instance, the importance of certain nodes and links in the network, which nodes are the most important and critical, which are the most central for passing on information, etc.
In a previous master dissertation the application of SNA to enterprise architecture (EA) has been successfully explored, for applications like change impact analysis (i.e., what is the impact of replacing some software application?), gap analysis (i.e., how can we define the migration path from the as-is architecture to a to-be architecture and what are the most critical components in this change process), and view definition (i.e., can the complexity of enterprise architecture models be reduced by further abstraction that is informed by the SNA analysis?).
The purpose of the new thesis is to continue the work that has started in different directions, including
– Application of SNA to EA frameworks with defined viewpoints. Instead of building a social network or graph for the entire EA model, we will explore SNA for different viewpoints that define views over the EA model. This way, SNA results will become more easier to interpret, hence potentially more useful.
– Matching existing EA analysis techniques with SNA analysis techniques. After the initial explorative study, a more in-depth comparison of existing techniques is needed, to identify further opportunities of SNA for EA analysis.
– Automation of SNA metrics and implementation of SNA techniques in an EA modeling tool (e.g., the Archi tool)
– Demonstration and evaluation of the potential of SNA for EA in case-studies (either existing documented case-studies or new cases)
If taken as a duo thesis, additional emphasis will be put on implementation aspects.
Reference