MVUC: An Interactive System for Mining and Visualizing Urban Co-locations

摘要

Spatial co-location patterns and rules may reveal the spatial associations between spatial features whose instances are frequently located in a spatial neighborhood. This paper develops an interactive system for Mining and Visualizing Urban Co-locations (MVUC) from the Point Of Interest (POI) datasets of cities. According to user-specified thresholds, MVUC efficiently mines maximal patterns based on instance-trees. Then MVUC concisely demonstrates the table-instances of patterns on map based on the convex hull algorithm, intuitively illustrates areas in which the consequents' instances of rules possibly appear on map based on the intersection of areas, and visually figures patterns by graph. MVUC also allows users to compare co-locations in different regions. MVUC can help users with analyzing the functional regions of cities, selecting the locations for public services, and other urban computing.