Neural Embeddings of Urban Big Data Reveal Emergent Structures in Cities

Abstract

Over decades, many cities have been expanded and functionally diversified by population activities, sociodemographics and attributes of the built environment. Urban expansion and development have led to emergence of spatial structures of cities. Uncovering cities' emergent structures and spatial gradients is critical to the understanding of various urban phenomena such as segregation, equality of access, and sustainability. In this study, we propose using a neural embedding model-graph neural network (GNN)-that leverages the heterogeneous features of urban areas and their interactions captured by human mobility network to obtain vector representations of these areas. Using large-scale high-resolution mobility data sets from millions of aggregated and anonymized mobile phone users in 16 metropolitan counties in the United States, we demonstrate that our embeddings encode complex relationships among features related to urban components (such as distribution of facilities) and population attributes and activities. The spatial gradient in each direction from city center to suburbs is measured using clustered representations and the shared characteristics among urban areas in the same cluster. Furthermore, we show that embeddings generated by a model trained on a different county can capture 50% to 60% of the emergent spatial structure in another county, allowing us to make cross-county comparisons in a quantitative way. Our GNN-based framework overcomes the limitations of previous methods used for examining spatial structures and is highly scalable. The findings reveal non-linear relationships among urban components and anisotropic spatial gradients in cities. Since the identified spatial structures and gradients capture the combined effects of various mechanisms, such as segregation, disparate facility distribution, and human mobility, the findings could help identify the limitations of the current city structure to inform planning decisions and policies. Also, the model and findings set the stage for a variety of research in urban planning, engineering and social science through integrated understanding of how the complex interactions between urban components and population activities and attributes shape the spatial structures in cities.

Hyper‐Local Temperature Prediction Using Detailed Urban Climate Informatics

“…For example, trees in the south of the node may have a more persistent cooling effect from shading than trees in the north. To preserve the spatial information within the averaging patch, more advanced spatial embedding techniques are desired in future studies (Fan et al, 2021).…”

Human Mobility Disproportionately Extends PM2.5 Emission Exposure for Low Income Populations

“…These methods from prior studies allow us to estimate fine-scale resolutions of PM 2.5 concentrations at individual residences. However, exposure to PM 2.5 emissions is also influenced by dynamic human activities-work and life activities-that occur mostly outside the home, creating variations in the exposure to ambient emissions [15,16].…”

“…Specifically, this contextual information enables the capture of frequently visited places of individuals and the length of stay at the places to uncover the daily life needs and work activities of city dwellers [30]. In short, the gyration radius and contextual information in analytics of mobile phone data have promoted a variety of applications, including the examination of disparities in disaster evacuation by race and income [17], quantification of income segregation during social interactions [31], and characterization of urban structures with social and demographic inequalities [16]. These studies [32,33,34], however, have not considered the intersection between human dynamics and the environment to understand how mobility influences the exposure of the populations to PM 2.5 air pollutants.…”

Specifying Evacuation Return and Home-switch Stability During Short-term Disaster Recovery Using Location-based Data

“…With location-based data, several studies have examined population mobility during disasters (Coleman, Gao, DeLeon, & Mostafavi, 2021;Gray & Mueller, 2012;Hsu, Fan, & Mostafavi, 2021;Lu, Bengtsson, & Holme, 2012;Pastor-Escuredo et al, 2014;Yabe, Ukkusuri, & C. Rao, 2019) and assessed disaster impacts (Bonaccorsi et al, 2020;Esmalian, Yuan, Xiao, & Mostafavi, 2022;Fan, Jiang, & Mostafavi, 2020;Lee, Maron, & Mostafavi, 2021;Q. Wang & Taylor, 2014;Yuan, Yang, Li, & Mostafavi, 2022); however, the majority of these studies focus on evacuation patterns (Deng et al, 2021;Song et al, 2016), disruption in mobility (Arrighi, Pregnolato, Dawson, & Castelli, 2019;Esmalian et al, 2021;Galeazzi et al, 2021), and mobility resilience (Fan, Jiang, & Mostafavi, 2021;Fan, Yang, & Mostafavi, 2021;Roy, Cebrian, & Hasan, 2019;Y. Wang, Wang, & Taylor, 2017).…”