Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Immersive Analytics of Anomalies in Multivariate Time Series Data with Proxy Interaction

: Kloiber, Simon; Suschnigg, Josef; Settgast, Volker; Schinko, Christoph; Weinzerl, Martin; Schreck, Tobias; Preiner, Reinhold


Sourin, Alexei (Ed.) ; Institute of Electrical and Electronics Engineers -IEEE-; European Association for Computer Graphics -EUROGRAPHICS-:
International Conference on Cyberworlds, CW 2020. Proceedings : 29 September - 1 October 2020, Caen, France, online
Los Alamitos, Calif.: IEEE Computer Society Conference Publishing Services (CPS), 2020
ISBN: 978-1-7281-6497-7
ISBN: 978-1-7281-6498-4
International Conference on Cyberworlds (CW) <19, 2020, Online>
Fraunhofer Austria ()
multivariate time series

In industry and science, sensor data play a vital role in research, optimisation, monitoring, testing and many other use cases. When performing tests with repeated cycles of similar behaviour, e.g., durability tests, it is often important to find anomalous sensor behaviour that deviates from regular patterns in the data. We here explore the design space of VRbased immersive analytics for time series data, for use e.g., in engineering contexts where an underlying application is also given in VR. The use of 3D visualisation for time series exploration is a much-discussed topic and careful consideration for its use must be taken. With the rise of immersive environments, we re-visit the classic problem of 3D time series visualisation and introduce an immersive walk-up usable interaction proxy that supports efficient navigation of otherwise possibly occluded time series views. The proxy indicates anomalies in the data for easy access and provides efficient zooming and filtering controls, among other effective interaction possibilities. This approach is combined with suitable data analysis techniques, providing an environment for effective and efficient immersive anomaly detection and comparative data analysis that we call WaveCharts. We demonstrate the applicability of our approach by two real-world use cases, and we discuss the necessary tools it provides to aid the analysis process of large sensor data.