An adapted centrality index to assess spatial accessibility in street networks: Application to two medium-sized cities in Brazil

This paper explores an adapted network-based measure for assessing spatial accessibility in urban street networks: the Potential Accessibility (PA) index. Building upon network centrality measures and incorporating concepts from transportation and geography studies, the proposed index integrates spatial competition for limited-capacity opportunities into a directed, weighted network model. The PA index accounts for both supply (e.g., employment opportunities) and demand (e.g., population) attributes, and their spatial interaction within a global analysis or a local one (with a predefined catchment radius). To demonstrate the PA index's properties and practical relevance, we apply it to two medium-sized Brazilian cities, analyzing accessibility to formal employment opportunities. We conducted four experiments: an unweighted analysis, a supply-weighted analysis, a competition-based analysis with both supply and demand weights, and a local accessibility analysis using a 1600-m radius. A fifth experiment compares the PA index to Reach centrality, a non-competitive network-based accessibility measure. Results reveal how variations in street network configuration influence accessibility outcomes and inequality patterns, captured using the Palma Ratio. The PA index distinguishes local from global accessibility more effectively than the Reach index and better reveals how competition and proximity affect opportunity access. The method's flexibility allows for multiscalar analysis, integration of social attributes, and scenario simulations, supporting planning applications aimed at equitable access to services. The study contributes to accessibility research especially by introducing a network-based framework with the potential to enhance proximity-centered and local accessibility urban policies, particularly relevant to small and medium-sized cities.