Spatial clustering in SOLAP systems to enhance map visualization

Abstract

The emergence of the SOLAP concept supports map visualization for improving data analysis, enhancing the decision making process. However, in this environment, maps can easily become cluttered losing the benefits that triggered the appearance of this concept. In order to overcome this problem, a post-processing model is proposed, which relies on Geovisual Analytics principles. Namely, it takes advantage from the user interaction and the spatial clustering approach in order to reduce the number of elements to be visualized when this number is inadequate to a proper map analysis. Moreover, a novel heuristic to identify the threshold value from which the clusters must be generated was developed. The proposed post-processing model takes into account the query performed, i.e., the number of spatial attributes, the number of spatial dimensions, and the type of spatial objects selected from dimensions. The results obtained so far show: (i) the novel approach to support queries with two spatial attributes from different dimensions allows useful analysis; (ii) the proposed post-processing model is very effective in maintaining a map suitable to the user’s cognitive process; and, (iii) the heuristic proposed provide the user participation in the clustering process, in a user-friendly way.