In medicine, a steadily growing data complexity often paralleling new developments in image and non-image data acquisition is observed. This complexity poses many challenges on data processing, visualization, and exploration. It renders the design of overview visualizations, which convey all interesting patterns contained in the data, impossible. Instead, automatic data analysis techniques must be tweaked based on expert knowledge and combined with interactive visualizations for the retrieval of such patterns. This approach is at the heart of the field visual analytics.
In this talk, I present visual analytics approaches to investigating complex data from clinical medicine and epidemiology. In particular, the talk's main topics are (1) the integration of data mining and interactive visualizations for the visual analysis of simulated cerebrovascular hemodynamic data, (2) a methodology for joint visual analytics of image and non-image epidemiological population study data, and (3) the visual verification of cancer staging for therapy decision support.