Conference paper examining forecasting of pedestrian link flows in multilevel pedestrian networks in four Singapore precincts (Orchard Road, Raffles Place, Jurong East, Tampines). The study draws 3D pedestrian networks (indoor/outdoor/vertical) from OpenStreetMap and building plans, attaches land-use and GFA data, conducts over 250 pedestrian counts in 2018–2019, computes accessibility, space syntax (integration/choice) and two-phase betweenness metrics using multiple distance definitions (metric, angular, perceived), and fits lasso/ridge regression models. Main findings: betweenness (choice) combined with land-use yields strong predictive power; angular measures often outperform metric/perceived in some sites; links between MRT stations and retail/commercial GFA attract most footfall; multi-level (indoor/private) links are important for accurate forecasts.
Questions this document answers
2022-08-12
Orchard Road; Raffles Place; Jurong East; Tampines
During 2018 and 2019 at over 250 locations (three consecutive days per site).