3D Charts – Taxonomy and Implementation in a Virtual Globe
ISBN 978-85-88783-11-9
Authors
1Schnürer, R.; 2Eichenberger, R.; 3Sieber, R.; 4Hurni, L.
1INSTITUTE OF CARTOGRAPHY AND GEOINFORMATION, ETH ZURICH Email: schnuerer@karto.baug.ethz.ch
2INSTITUTE OF CARTOGRAPHY AND GEOINFORMATION, ETH ZURICH
3INSTITUTE OF CARTOGRAPHY AND GEOINFORMATION, ETH ZURICH
4INSTITUTE OF CARTOGRAPHY AND GEOINFORMATION, ETH ZURICH
Abstract
High-performance computers and advancements in computer graphics fostered the development of 3D cartography. Great progress has been made in creating digital 3D city models and implementing virtual globe engines to visualize topographic and thematic geo-data. In both cases however, 3D charts are used quite rarely. 3D city models often map the information on the objects themselves, whereas virtual globes’ capabilities to compose multi-parametric 3D objects are very limited. One of the more recent attempts towards 3D charts is made by Microsoft's Power Map for Excel where stacked 3D bars and extruded pie charts can be created on a virtual globe surface. From a cartographic point of view however, the number of different chart types and customization possibilities is rather small. The potential for representing data by including the third dimension is also not fully tapped. Hence, for the future development of this and other virtual globe engines, a concise framework for 3D charts is essential for software engineers as well as for map editors. We therefore propose a taxonomy for 3D charts which extends existing chart classifications for 2D maps (e.g. Schnabel 2007). Charts can be distinguished according to their geometry type into 3D point, line, surface, and solid charts. While a few special cases exist for points, lines, and surfaces (e.g. 3D areas), shapes may vary for solids (e.g. spheres, cylinders, stars). Sizes of solids can be scaled proportionally to attribute values by changing the volume, by adjusting the angle, by duplication, or by extrusion. In case of multivariate data, solids can be partitioned or replicated by the number of attributes. Segments or replicated solids may be aligned differently (e.g. in a grid, along a spiral line) and properties of alignments may change (e.g. gaps, anchor points, or ordering). Lastly, different chart types can be combined, for example by placing 3D bars into donuts to create Abacus-like charts. The proposed taxonomy might serve as a reference guide for future usability studies examining the feasibility between 2D and 3D charts. Moreover, it will help to develop animation concepts when querying attribute values and comparing them between charts. As not all charts will be suitable for use in practice, we align the taxonomy with concepts of the 3D Mapping Space (Sieber et al. 2013). It is for instance desirable to recognize charts from a bird’s eye view and gain additional information when tilting the globe. Based on these concepts, we implement a series of 3D charts and insert them into the virtual globe toolkit osgEarth. Charts are defined in JavaScript configuration files, generated with OpenJSCAD, adjusted with Blender, and displayed as 3D geometries in the virtual globe. With this workflow, we are able to create stacked pyramid frustums, 3D coin charts, nested hemispheres, and helix charts, amongst others. The underlying data originates from the Atlas of Switzerland and thus serves as a realistic test set for proving the feasibility of the workflow. The JavaScript code for creating charts may be reused in other virtual globes - e.g. Cesium, Google Earth - and for other cartographic products. By implementing novel 3D chart types, it is intended to inspire cartographers and to provide attractive but at the same time still readable maps for experts and the general public.