摘要
Recently,the expertise accumulated in the field of geovisualization has found application in the visualization of abstract multidimensional data,on the basis of methods called spatialization methods.Spatialization methods aim at visualizing multidimensional data into low-dimensional representational spaces by making use of spatial metaphors and applying dimension reduction techniques.Spatial metaphors are able to provide a metaphoric framework for the visualization of information at different levels of granularity.The present paper makes an investigation on how the issue of granularity is handled in the context of representative examples of spatialization methods.Furthermore,this paper introduces the prototyping tool Geo-Scape,which provides an interactive spatialization environment for representing and exploring multidimensional data at different levels of granularity,by making use of a kernel density estimation technique and on the landscape "smoothness" metaphor.A demonstration scenario is presented next to show how Geo-Scape helps to discover knowledge into a large set of data,by grouping them into meaningful clusters on the basis of a similarity measure and organizing them at different levels of granularity.
Recently, the expertise accumulated in the field of geovisualization has found application in the visualization of abstract multidimensional data, on the basis of methods called spatialization methods. Spatialization methods aim at visualizing multidi- mensional data into low-dimensional representational spaces by making use of spatial metaphors and applying dimension reduction techniques. Spatial metaphors are able to provide a metaphoric framework for the visualization of information at different levels of granularity. The present paper makes an investigation on how the issue of granularity is handled in the context of representative examples of spatialization methods. Furthermore, this paper introduces the prototyping tool Geo-Scape, which provides an interactive spatialization environment for representing and exploring multidimensional data at different levels of granularity, by making use of a kernel density estimation technique and on the landscape ＂smoothness＂ metaphor. A demonstration scenario is presented next to show how Geo-Scape helps to discover knowledge into a large set of data, by grouping them into meaningful clusters on the basis of a similarity measure and organizing them at different levels of granularity.
