Forecasting district-wide pedestrian volumes in multi-level networks in high-density mixed-use areas

Loading...
Thumbnail Image
Authors
Mavros, Panos
Helle, Veera
Acebillo, Pablo
Xu, Shuchen
Author (Corporation)
Publication date
06.2022
Typ of student thesis
Course of study
Type
04B - Conference paper
Editors
van Nees, Akkelies
de Koning, Remco Elric
Jacobsen Åsli, Thale
Editor (Corporation)
Supervisor
Parent work
13th International Space Syntax Symposium
Special issue
DOI of the original publication
Link
Series
Series number
Volume
Issue / Number
Pages / Duration
Patent number
Publisher / Publishing institution
Western Norway University of Applied Sciences (HVL)
Place of publication / Event location
Bergen
Edition
Version
Programming language
Assignee
Practice partner / Client
Abstract
This paper is concerned with improvements in the forecasting of pedestrian flows in multilevel pedestrian networks in high-density urban environments. 3D network topology measures are combined with land-use data, and validated against extensive pedestrian counts, to provide both evidence for the applicability of network analysis in tropical metropolises, as well as a calibrated tool for urban planners. The research focuses on four area in Singapore. These areas have in common that they all are prominent transport hubs, but differ in surrounding land-use types and dominant network topology (e.g. indoor, outdoor, above ground, below ground, at grade). Multi-level pedestrian networks were drawn based on OpenStreetMap, include sidewalks on both sides of major roads for a radius up to 2 kilometres from the site centroids. Spatial network analysis was performed using sDNA which allows vertical networks to generate measures describing the spatial configuration of the network. Subsequently, pedestrian counts were conducted during three consecutive days. In total, counts were conducted at more than 250 locations in 2018 and 2019, well before the global COVID19 pandemic. Pedestrian flows are set against a series of variables, including pedestrian attractors and generators (e.g. shops, offices, hotels, dwellings), and variables describing the spatial configuration of the network, using advanced regression models. Our results show that betweenness metrics (i.e. space syntax choice) combined with land-use yield high predictive power. Dependent on the study site, network metrics based on angular distance outperform those based on metric distance or perceived link distance. This research demonstrates that is necessary to account for the multi-level nature of networks, and that indoor flows through private developments cannot be neglected, in particular when planning for integrated transport developments. The paper concludes with recommendations and implications for practice.
Keywords
Pedestrian volume, Accessibility, Pedestrian demand, Space Syntax, sDNA, Multi-level network
Project
Event
13th International Space Syntax Symposium
Exhibition start date
Exhibition end date
Conference start date
20.06.2022
Conference end date
24.06.2022
Date of the last check
ISBN
ISSN
Language
English
Created during FHNW affiliation
Yes
Strategic action fields FHNW
Publication status
Preprint
Review
Peer review of the complete publication
Open access category
Closed
License
Citation
Mavros, P., van Eggermond, M., Erath, A., Helle, V., Acebillo, P., & Xu, S. (2022). Forecasting district-wide pedestrian volumes in multi-level networks in high-density mixed-use areas. In A. van Nees, R. E. de Koning, & T. Jacobsen Åsli (Eds.), 13th International Space Syntax Symposium. Western Norway University of Applied Sciences (HVL). https://doi.org/10.26041/fhnw-4153