This paper advances conventional OD estimation models from cross-sectional counts by introducing an Origin–Stopover–Destination (OSD) framework that explicitly models time-varying in-store dwell between origins and destinations. We integrate cross-sectional flow surveys, mobile phone location data, and occupant observations to estimate unknown parameters. Applied to a large underground pedestrian mall, the model estimates inflows/outflows, aisle-level flows, and store occupancy, yielding close agreement with observations. Inferred occupancies reveal attribute-dependent mean dwell times across stores. We further simulate traffic shifts under a subway extension scenario.
Osaragi et al. (Sun,) studied this question.