Tong, ChaoRoberts, RichardLaramee, Robert S.Berridge, DamonThayer, DanielTao Ruan Wan and Franck Vidal2017-09-212017-09-212017978-3-03868-050-5https://doi.org/10.2312/cgvc.20171276https://diglib.eg.org:443/handle/10.2312/cgvc20171276The National healthcare Service (NHS) in the UK collects a massive amount of high-dimensional, region-centric data concerning individual healthcare units throughout Great Britain. It is challenging to visually couple the large number of multivariate attributes about each region unit together with the geo-spatial location of the clinical practices for visual exploration, analysis, and comparison. We present a novel multivariate visualization we call a cartographic treemap that attempts to combine the space-filling advantages of treemaps for the display of hierarchical, multivariate data together with the relative geo-spatial location of NHS practices in the form of a modified cartogram. It offers both space filling and geospatial error metrics that provide the user with interactive control over the space-filling versus geographic error trade-off. The result is a visualization that offers users a more space efficient overview of the complex, multivariate healthcare data coupled with the relative geo-spatial location of each practice to enable and facilitate exploration, analysis, and comparison. We evaluate the two metrics and demonstrate the use of our approach on real, large high-dimensional NHS data and derive a number of multivariate observations based on healthcare in the UK as a result. We report the reaction of our software from two domain experts in health science.10.2312/cgvc.2017127629-42