Am Freitag 14.12.2018 um 12 Uhr findet die Verteidigung der Dissertation von DI Christian Partl in der Inffeldgasse 16b, Seminarraum IGI, 1. Stock, Raum 074 statt.
Interessierte Personen sind herzlich zur Teilnahme eingeladen.
Abstract der Dissertation:
Multivariate graphs are commonplace in many application domains, such as social sciences, transport, and molecular biology. The analysis of such graphs can be challenging due to large quantities of data, the complexity of relationships, and data heterogeneity. However, visualization can help humans to gain insights into this complex data.
In this thesis, we introduce four novel visualization techniques for multivariate graphs: EnRoute, Entourage, Pathfinder, and ConTour. Paths play a primary role in all four visualization techniques. Either, paths are utilized to alleviate graph exploration, or the analysis of paths is the designated goal. EnRoute is a scalable visualization for the investigation of graph attributes, where paths can be interactively extracted from a graph to show them side-by-side with large amounts of heterogeneous attributes. Entourage allows users to investigate multiple graph partitions simultaneously by showing one partition in detail, while showing only contextually relevant paths for other partitions.
Pathfinder is a query-driven technique that uses ranking and alternative path representations for the analysis of multiple paths in large graphs.
Finally, ConTour allows users to investigate path relationships in heterogeneous linked datasets in a list-based interface. While the core concepts of all proposed visualization techniques are domain-agnostic, the visualizations were primarily developed to help experts to investigate multivariate graphs in molecular biology and drug discovery.
Therefore, our techniques are evaluated in case studies and usage scenarios that illustrate their fitness to support domain experts with their analysis tasks.
