Urban network parcellation using graph theoretic approach

Abstract

This paper investigates the street-based network topology of urban spaces as represented in Geographic Information Systems (GIS) databases. We apply graph theoretical spectral methods to analysis of ways in maps. Our findings show that such methods can find natural parcellations of spaces, thus capturing the significance and semantics of a location from analysis of topological structure of the underlying map connectivity structure. Such analysis might be important for semantic search in the “5D World Map” project.