A heuristic alpha-shape based clustering method for ranked radial pattern data

Existing point clustering methods are not specifically designed for ranked radial pattern data. This article presents a novel method for summarizing and visualizing interactions between an origin and a number of destinations. The method is driven to bridge the methodological gap to design a point clustering method to maintain both origin–destination direction and ranking information for those data. Based on α-shape, an established concept and measure in computational geometry, the minimum bounding hull (MBH) is defined and utilized. MBH is chosen to represent ranked radial pattern point data and a heuristic α-shape based (HAS) clustering method is designed. HAS shows advantages to preserve both spatial compactness and attributive homogeneity. The case study of Zhengzhou, China’s outgoing telephone records sheds lights on potential applications of MBH and HAS methods for geographic data. The innovation of this article lies in two aspects: using α-shape, a computational geometry metric heuristically and clustering for ranked radial pattern data.

Research highlights
► A novel method for summarizing and visualizing interactions between an origin and a number of destinations, ranked radial pattern geographic data.
► Alpha-shape, a computational geometry metric is used heuristically for the clustering method.
► The case study of Zhengzhou, China’s outgoing telephone records sheds lights on potential applications of the method.