Analysis of the spatial relations among point distributions on a discrete space

This paper proposes a method of analyzing spatial relations among point distributions on a discrete space. Spatial topology and proximity are discussed in an integrated framework at both local and global scales. Local relations are described by geographical representations, whereas global relations are visualized by graph representations. The latter also provide a means of classifying the point distributions. The proposed method is applied to school location planning in Japan. The results reveal the appealing properties of the method and provide empirical findings.