Masterarbeit
Visual Exploration of Entity-Level Changes in Knowledge Graphs
Completion
2026/08
Research Area
Intelligent Information Management
Students
Roshita Shakya
Advisers
Christoph Göpfert M.Sc.
Description
The objective of this master’s thesis project is to make changes in knowledge graphs easier to track and understand by developing an intuitive visualization method. Many current solutions track changes through logs of queries issued to the triplestore, but interpreting such edit histories often requires technical expertise and is not easily accessible to non-technical users. Furthermore, they are typically displayed in tabular views, listing changes on the graph-level. This can become overwhelming to manage, especially for larger knowledge graphs or when changes occur frequently. As a result, these approaches are not sufficiently user-friendly and do not enable an easy way to inspect changes at the attribute level for specific entities.
First, a requirements analysis must be conducted, and a review of the state of the art in visualizing changes in RDF data must be conducted. The identified approaches then need to be classified and evaluated according to the requirements. Based on these findings, an approach to visually explore modifications to specific entities in the knowledge graph needs to be conceptualized. To achieve this, a shape-based approach should be explored. Shapes can be used to represent the structure of entities. As part of this project, it should be analyzed to what extent they are suitable for restricting visualization to relevant subgraphs. The conceptual feasibility of the proposed visualization approach should be demonstrated through a prototypical implementation. The effectiveness and usability of the solution should then be assessed in a suitable evaluation, with a particular focus on how well users can understand and track changes in the knowledge graph.
