Beyond Binary Connections: Asymmetry and Weighted Isovist Networks in Advanced VGA Analysis

Visibility graph analysis (VGA) is a mainstay research tool used in the field of space syntax. However, there are many questions regarding the explanatory power of VGA that have yet to be explored. Our work begins with the observation that current VGA relies on an underlying Boolean graph for the computation of configurational space syntax measures. In this paper, we explore a more nuanced approach whereby the relative area of intersection between two isovists is incorporated into a weighted network. We demonstrate that this new approach, utilising weighted and directed visibility graphs, is capable of incorporating and representing more information about the connection between two isovists. It can also represent certain asymmetric qualities of the connection, which has not typically been a feature of traditional space syntax analysis. In this paper we introduce weighted and directed visibility graphs, discussing the differences and similarities with traditional VGA. Finally, we demonstrate that by adding weights, based on area of isovist overlap, to a visibility graph, we improved its correlation with pedestrian counts by a notable 32%. Our paper introduces an innovative approach to VGA, emphasising isovist overlap as a weighted factor, significantly enhancing the correlation with observed movement patterns. This opens new avenues for understanding urban and building systems, although further research is essential to comprehensively explore the underlying mechanisms.