Multivariate-Space-Time Cube to Visualize Multivariate Data

要旨

The data visualization refers to the representation of the data tables with the techniques of computer graphics to enable users to draw the information implicit in data. The multivariable spatio-temporal data consist of the variables of time, location, and attributes that are represented on tables as data fields. The article approaches 3-dimensional Cartesian coordinates to representing adequately the data of a table. The main idea is that each record of the data table is converted to a data plane of two dimensions, i.e. each data tuple of a record is shown as the graphics in a 2-dimensional domain. Then these data planes are arranged along and perpendicularly to the time axis of a 3-dimensional Cartesian coordinate system. This method includes the data of attributes in a space-time cube (STC) to represent adequately the data fields of a table. This approach is called multivariate-space-time cube (MSTC). The multivariate-space-time cube proposed in this paper not only represents visually and adequately the data of a table, but also offers users straight views on non-inclined maps of geographic area. The proposed multivariate-space-time cube is illustrated with the data table of Napoleon’s 1812 Russian campaign. All data of the table is represented on a cube showing visually the great loss on soldiers as well as tragic defeat of the campaign.