Graph Simplification Solutions to the Street Intersection Miscount Problem

Street intersection counts and densities are fundamental measures in transport geography and urban planning. Conventional street network data and analysis tools often lead to significant overcounting of these measures. This study investigates the causes of such overcounting and proposes remedies. It introduces algorithms to streamline urban street network models through edge simplification and node consolidation, which enhance computational efficiency and improve accuracy in network measures like intersection counts, street segment lengths, and node degrees. The algorithms are validated and subjected to a global empirical assessment to evaluate the extent of count bias. These findings highlight the prevalence of this bias and underscore the necessity for better methods to mitigate inaccuracies in intersection representation.