Visualising large semantic datasets

• Main Author: Mazumdar, Suvodeep
• Other Authors: Ciravegna, Fabio ; Petrelli, Daniela
• Published: University of Sheffield 2013
• Subjects: 004
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.605324

This thesis aims at addressing a major issue in Semantic Web and organisational Knowledge Management: consuming large scale semantic data in a generic, scalable and pleasing manner. It proposes two solutions by de-constructing the issue into two sub problems: how can large semantic result sets be presented to users; and how can large semantic datasets be explored and queried. The first proposed solution is a dashboard-based multi-visualisation approach to present simultaneous views over different facets of the data. Challenges imposed by existing technology infrastructure resulted in the development of a set of design guidelines. These guidelines and lessons learnt from the development of the approach is the first contribution of this thesis. The next stage of research initiated with the formulation of design principles from aesthetic design, Visual Analytics and Semantic Web principles derived from the literature. These principles provide guidelines to developers for building generic visualisation solutions for large scale semantic data and constitute the next contribution of the thesis. The second proposed solution is an interactive node-link visualisation approach that presents semantic concepts and their relations enriched with statistics of the underlying data. This solution was developed with an explicit attention to the proposed design principles. The two solutions exploit basic rules and templates to translate low level user interactions into high level intents, and subsequently into formal queries in a generic manner. These translation rules and templates that enable generic exploration of large scale semantic data constitute the third contribution of the thesis. An iterative User-Centered Design methodology, with the active participation of nearly a hundred users including knowledge workers, managers, engineers, researchers and students over the duration of the research was employed to develop both solutions. The fourth contribution of this thesis is an argument for the continued active participation and involvement of all user communities to ensure the development of a highly effective, intuitive and appreciated solution.