Search

This is a follow-up of our previous study (Investigate visualizations that wrap around spherical topologies - https://osf.io/vctfu). The difference with the previous study are: - We look at node-link representation of networks that wrap around toroidal and spherical topologies. - We chose two spherical layouts inspired from the previous study: hemisphere (Orthographic hemisphere) and non-hemisphere (Equal Earth) projections. - We compare them with existing toroidal layouts and regular flat (non-spherical, non-toroidal) representations. - All the techniques support interactive panning/rotation except the regular node-link layout. This allows users to view the continuation of edge wrapping around the display

Traditional 2-dimensional (2D) data visualisations are designed to be viewed on a printed page or standard screen. However, it is possible to arrange data on a 3D surface such as a sphere or torus topology, and then project such a surface onto a 2D rectangular plane while wrapping the data around the boundaries of a 2D surface. Potential benefits include the preservation of the data’s underlying topology in the case of geospatial data, and the preservation of the data’s underlying structure in the case of abstract data such as networks. However, in wrapped visualisations the viewer needs to perceive continuity of the visualisation across boundaries, which may induce additional cognitive load. Despite widespread use of projection techniques in geovisualisation and study of network embeddings on surfaces by mathematicians, there is very little research into wrapped data visualisations. This research aims to (1) explore the design space for interactive wrapped data visualisations and determine the kinds of data visualisations that can potentially benefit from being laid out on a 2D cylindrical, spherical, or toroidal topology; (2) design a tool capable of creating wrapped visualisations for various types of data, which can be interactively panned; (3) conduct user studies to evaluate readability of wrapped visualisations and their equivalent unwrapped visualisations; (4) conduct user studies to evaluate whether interactive wrapped visualisations for real-world applications such as software applications could be practically usable by domain experts.

Traditional 2-dimensional (2D) data visualisations are designed to be viewed on a printed page or standard screen. However, it is possible to arrange data on a 3D surface such as a sphere or torus topology, and then project such a surface onto a 2D rectangular plane while wrapping the data around the boundaries of a 2D surface. Potential benefits include the preservation of the data’s underlying topology in the case of geospatial data, and the preservation of the data’s underlying structure in the case of abstract data such as networks. However, in wrapped visualisations the viewer needs to perceive continuity of the visualisation across boundaries, which may induce additional cognitive load. Despite widespread use of projection techniques in geovisualisation and study of network embeddings on toroidal surfaces, there is very little research into wrapped data visualisations. This project aims to (1) conduct user studies to evaluate readability of interactively wrapping on map projection visualisations, and compare with their equivalent unwrapped visualisations; (2) conduct user studies to evaluate the readability of interactive wrapped visualisations of spherical node-link layouts and compare with their toroidal representations.

investigate torus drawing for network visualization