Exploring 3D spatial morphology using multilayered space syntax, network science and wi-fi log data

Spatial morphology of 2D space has been well studied at the scales of building, community, and city in space syntax literature. Space syntax decomposes continuous geographic space into a set of unique axial lines and represents them as a spatial network to analyze spatial morphology. However, 3D spatial morphology remains largely unexplored, partially due to limited data access and methodological constraints. Motivated by the multilayered network literature in network science, this work extends the conventional 2D space syntax axial network into a multilayered axial network to explore the spatial morphology of a university campus. Scaling properties are discovered from several aspects of the constructed spatial network (e.g., degree, local integration). The relationship between spatial morphology and route-based pedestrian flow is evaluated using a large Wi-Fi log dataset and multiple regression analysis. We found a significant correlation between the two, with a correlation coefficient of 0.756 and r² of 0.571. Four network features (i.e., local depth, geometric length, eigenvector, and betweenness), are found to significantly shape the pedestrian flow. The results of community detection show the effectiveness of multilayered space syntax analysis in depicting functional areas of campus, despite the complex nature of 3D space. Whilst a limitation of this research is the need to account for the 3D angular route preferences when characterizing the morphology of campus and the flow estimated based on the shortest routes using the Dijkstra algorithm.