USING COMPLEX NETWORK ANALYSIS AND VISUALISATION TO ANALYSE PROBLEMATIC ENTERPRISE SCALE INFORMATION SYSTEMS?

David Greenwood; Ian Sommerville

Authors

David Greenwood
Ian Sommerville

Keywords:

Sociotechnical Systems Engineering, Complex Network Analysis, Troubleshooting Information System Deployments

Abstract

Society is demanding larger and more complex information systems to support increasingly complex and critical organisational work. During or after the deployment of these systems it is typical for problematic socio-technical issues to arise. Whilst troubleshooting socio-technical issues in small-to-medium scale situations may be achievable using approaches such as rapid ethnography combined with a theoretical framework such as distributed cognition or activity theory; troubleshooting enterprise scale situations is an open research question because of the overwhelming number of socio-technical elements and interactions involved.

Techniques and tools for complex network analysis enable the analysis of systems comprising large numbers of nodes. These tools are becoming increasingly accessible to non-computer scientists and mathematicians and so have been used to analyse a diverse variety of large-scale systems from social networks through to metabolic pathways in living organisms. We believe there is scope to use similar techniques to facilitate the analysis of problematic enterprise scale socio-technical systems.

This paper demonstrates, via means of a case study, proof-of-concept tools for large-scale network analysis and visualisation that may provide a promising avenue for identifying problematic elements and interactions amongst an overwhelming number of socio- technical elements. We demonstrate the potential of this approach by showing that: i) a problematic situation may be represented as a directed graph such that the elements in the situation are represented as nodes, and interactions between nodes as edges; ii) that eigenvector centrality may be used to rank the importance of elements in a situation and that highly ranked elements match those identified as important by a human analyst; iii) the ‘complexity’ of a situation, or a part of a situation, may be characterised using a feedback degree score which provides an indication of the extent elements are highly interconnected and involved in feedback loops. These findings indicate that computers may be used to aid the analysis of problematic large-scale complex socio-technical situations by highlighting elements, or groups of interacting elements, that are important to the overall outcome of a problematic situation.

Author Biographies

David Greenwood

David Greenwood is a PhD student at the University of St Andrews funded by an LSCITS.ORG EPSRC studentship. David has a Master's degree in Applied Informatics from the University of Reading and holds an MA and BA in Natural Sciences from the University of Cambridge. He is a member of the UK Association for Information Systems and a member of the ACM. David's previous research has been in the interoperability and alignment of IT systems and business processes. Today his research focuses on a related area, the effect of social and organisational complexity on socio-technical systems.

Ian Sommerville

Ian Sommerville has been a professor of computer science at St Andrews since 2006 and was previously at Lancaster University. His research interests are primarily in complex systems engineering with a focus on dependability, requirements engineering and socio-technical systems. While at Lancaster, he cooperated with sociologists to study complex computer-based systems with a view to understanding the realities of their use and this has led to a long-term interdisciplinary collaboration. He is convinced that by examining social, organisational and human issues that we can build systems that offer faster 'time to value' after they have been deployed. His goal now is to make socio-technical systems engineering a reality where we use our understanding of socio-technical issues in the development process to create more usable and dependable software systems.

USING COMPLEX NETWORK ANALYSIS AND VISUALISATION TO ANALYSE PROBLEMATIC ENTERPRISE SCALE INFORMATION SYSTEMS?

Authors

Keywords:

Abstract

Author Biographies

David Greenwood

Ian Sommerville

Published

How to Cite

Issue

Section