Article Street network analysis ‘‘edge effects’’: Examining the sensitivity of centrality measures to boundary conditions Jorge Gil Department of Architecture, Chalmers University of Technology, Gothenburg, Sweden Abstract With increased interest in the use of network analysis to study the urban and regional environment, it is important to understand the sensitivity of centrality analysis results to the so-called ‘‘edge effect’’. Most street network models have artificial boundaries, and there are principles that can be applied to minimise or eliminate the effect of the boundary condition. However, the extent of this impact has not been systematically studied and remains little understood. In this article we present an empirical study on the impact of different network model boundaries on the results of closeness and betweenness centrality analysis of street networks. The results demonstrate that the centrality measures are affected differently by the edge effect, and that the same centrality measure is affected differently depending on the type of network distance used. These results highlight the importance, in any study of street networks, of defining the network’s boundary in a way that is relevant to the research question, and of selecting appropriate analysis parameters and statistics. Keywords Distance metrics, model validation, network graphs, network structure, street networks Introduction Spatial network analysis is an increasingly used method in the study of the urban and regional built environment, in order to address a wide range of phenomena, including human mobility, economic and energetic performance, social equity, individual health, and sustainable development in general. Empirical studies dealing with real world spatial networks use in most cases artificially bounded network models. These network models exclude network elements and events occurring beyond the boundary of the model and, because the analytic algorithms are fundamentally relational, such an omission can affect Corresponding author: Jorge Gil, Department of Architecture, Chalmers University of Technology, SE-41296 Gothenburg, Sweden. Email: jorge.gil@chalmers.se Environment and Planning B: Planning and Design 0(0) 1–18 ! The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0265813516650678 epb.sagepub.com