Daisuke Hasegawa

This study examines how spatial configurations, institutional arrangements, and urban locations shape face-to-face interaction networks in innovation hubs. Two incubation facilities located in the central business district (CBD) area of a regional Japanese city were analyzed using long-term proximity sensing based on Bluetooth low-energy (BLE) beacons. The facilities differ in their spatial layout and management structure; one is an open, loosely partitioned, free-access space, whereas the other is a segmented, fee-based, and institutionally managed facility. Interaction networks were constructed by defining a tie as co-presence in the same area for 10 minutes or longer. The open facility exhibited a larger and more dispersed network, characterized by shorter interaction durations, whereas the managed facility showed a longer average interaction time per session (62.77 minutes compared to 46.79 minutes) and a more centralized structure associated with strong ties. Location effects were also observed. Although both facilities are embedded in a highly accessible CBD area, their position at a moderate distance from dominant universities enabled cross-organizational exchange. Cross-occupation interactions accounted for more than half of all ties in both facilities (56.7% and 61.2%), indicating that spatial neutrality facilitated bridging among heterogeneous actors. The results demonstrate that interaction networks vary systematically according to spatial and institutional design and that proximity-based sensing can provide empirical quantification of network assets at the district scale. Therefore, integrating spatial design, institutional frameworks, and location strategies is essential for planning innovation hubs that foster diverse and productive interactions.

1. 1.1. This paper addresses a new variant of the inverse shortest path problem (ISPP), aiming to infer implicit link costs from observed paths and detect outliers that deviate from consistent netw...