Measuring Proximity in a Graph of Spatial Data (ZIP Codes)

摘要

Proximity is used for showing closeness between objects. It is widely used in a social network, marketing, and online businesses. Proximity is also used for spatial analysis such as epidemic spread, extent of trade areas, and socioeconomic planning. Proximity in spatial data is always associated with distance. However, distance itself can be ambiguous when there is an impassable natural landmark barrier between them such as mountain and big lake. Impassable natural landmark barrier is not included in the boundaries of a ZIP Code. Adjacency between ZIP Codes would not happen if they did not have common boundaries. We propose a novel way of measuring proximity in a graph of spatial data based on common boundary points, centroid distance, and shortest path. By considering these three, we bound to get more exact precision answer for proximity and eliminating the possibility of ambiguous distance. We also provide a mathematical model for effectively calculating proximity by using the previously mentioned features.