A comparative assessment of edge-effect with syntax integration generated in axial and unit-segment approaches to modelling vehicular movement networks

Abstract

Unit-segment analysis – a recent development in space syntax – has shown better results in modelling vehicular movement networks than the traditional axial analysis merely by using configurational measures, integration results, of roadway segments of urban grids. Despite this advancement, some of the findings of the unit-segment model have remained controversial especially in the academic community. The concern of edge-effect is one of them. That is, at the practical level, how is it possible for a unit segment analysis to predict vehicular movement of a specified urban roadway grid with high accuracy when the analysis model does not consider trips that enter into the grid from outside? This paper throws a deeper insight into this question by evaluating the intensities of edge-effect generated by both the syntax models. By using the case of Tech Terrace, a residential neighbourhood of the city of Lubbock in West Texas, and then the city itself, the study produces extensive theoretical as well as empirical evidence showing that vehicular traffic predictions made by both the models are not free from edge-effect, but the generated effect in the unit-segment model is somewhat reduced than that of the axial counterpart. The findings also suggest that the problem of edge-effect is deep founded in the perception of network topology of an urban grid, and without understanding its role in assigning trips, brought from outside the grid into the grid itself, we indeed cannot understand the space syntax approach to modelling vehicular movement networks in a comprehensive way.