Spatial Structure Analysis Using Space Syntax Supported by Generative AI Interpretation

Abstract

This study examines the spatial structure of the cities of Zliten and Tarhuna through the integration of Space Syntax indicators with cognitive analysis supported by generative artificial intelligence. The research addresses the limited application of quantitative spatial analysis tools in Libyan urban studies, despite their importance in understanding movement networks and spatial relationships within cities. The study aims to develop an analytical framework that combines digital measurement with cognitive interpretation in order to support urban planning decisions. The research adopts a descriptive-analytical methodology based on axial analysis using the Depthmap software, alongside the utilization of generative language models for the interpretation and analysis of spatial outcomes. The findings reveal clear spatial differences between the two cities, The spatial analysis showed a mean Global Integration value of 0.67 and a mean Connectivity value of 3.32. Furthermore, the Choice index along the primary urban axes exceeded that of the secondary street network by more than 42%, highlighting the superior movement potential and strategic importance of the city's main corridors. particularly regarding intelligibility and cognitive connectivity within the urban networks. The study concludes by emphasizing the need to enhance urban connectivity and to develop digital analytical tools capable of supporting urban planning practices in Libyan cities. The originality of this study lies in the integration of Space Syntax indicators with explainable artificial intelligence (XAI) to establish a novel analytical framework capable of providing deeper and more interpretable insights into spatial relationships than those offered by conventional spatial analysis methods.