Seminar Computer Science - Steffen Frey

Title: Visual Analysis of Large Spatio-Temporal Data.

Abstract:

Advances in parallel computing systems, simulation methods and digital imaging lead to the generation of large and complex spatial data at an unprecedented rate. In this talk, I will discuss efficient visualization approaches that allow to gain insights from this data. To enable interactive user exploration, I will first outline the optimization and evaluation of visualization performance, with an emphasis on volume raycasting. Then, I will address in situ visualization which is an emerging paradigm to process data as it is generated. The visual analysis of complex data further requires an expressive presentation of aspects of interest. For this, I will discuss domain-driven approaches specifically designed to tackle research questions in different scientific fields as well as generic data-driven methods. In the second part of the talk, I will focus on the latter in detail and present visualization techniques using data similarity as the basis for analyzing spatio-temporal processes. In this context, different approaches for quantifying similarity and generating informative visualizations will be discussed, including the usage of machine learning models in the analysis. The talk is concluded with an outlook on addressing challenges in the visual analysis of increasingly large and complex data.