Resumen
Understanding the impact of transport systems on the spatial agglomeration of urban facilities is critical for urban and transport planning. Recent studies show three separate mechanisms, including matching, sharing, and trip chaining on the agglomeration of commercial facilities, but little is known about which of these mechanisms is dominant and how its dominance varies across transport systems. Aiming at empirically investigating the mechanisms, we first calculate a simple agglomeration index for 69 Japanese cities and then explore the association between the index and city-level characteristics (including transport) using a decision tree analysis. The results confirm that (1) cities with larger areas and higher train shares experience agglomeration, presumably through matching and/or trip chaining, while cities with smaller areas have less agglomeration despite high train shares; and (2) car-dependent cities experience agglomeration, presumably through sharing, particularly by agglomerating in their residential and roadside areas. These findings indicate that effective agglomeration forces vary across transport systems.