Multidimensional Projections are frequently employed to allow the layout, in lower dimensions (2D or 3D), of multidimensional objetcts, to support analysis based on a chosen criterion, such as similarity. Typically data sets ar abstract of unstructured and the result of a projection is a set of points mapped on a plane, where proximity suggests similarity. An alternative form of visual representation of such data is through similarity trees, where a hierarchy is imposed over the data points, based o the same criterion as the projection. This project proposes to develop new forms of representation of the results of these techniques, usually based on small icons. The work continues efforts in multidimensional visualization inside the VICG group, ICMC, with large applicability in various areas of knowledge. Additionally, it starts a new venue in the group, that of new designs for multidimensional layouts to improve their power of representation and visual appeal.
News published in Agência FAPESP Newsletter about the scholarship: