Abstract
Introduction
City is the ‘place’ in the ‘space’. Uncovering the interplay between physical and social configurations is essential and has witnessed a recent surge of interest from both the theoretical and practical sides of urban studies, planning and design. As Shaw et al. (2016: 1688) state, “there is a long tradition in linking the physical and relational spaces in the built environment. The development of powerful computing technology, emerging big and open data of interactions and flows, and theoretical perspectives on social-spatial processes has revolutionized the way in which we investigate social-spatial interactions”.
The development of such studies has evolved dramatically in conjunction with the advancement in information and communication technology (Liu et al., 2015; Wang et al., 2018). Nevertheless, existing efforts are largely made separately on this issue with the focus on either spatial or social patterns, which are now increasingly challenged by the emerging theoretical compositions, data resources, study scales and analytical methods. This special issue, therefore, aims to bring possible fruitful theoretical and practical implications through bridging the novel thoughts of strong conceptual overlaps on both physical and social aspects of the built environment. In particular, papers in this special issue address how social connections between individuals are shaping and are (re)shaped by the physical space attached to individuals. We hope that these discussions can contribute to place-making and the related knowledge translation between social science and spatial analysis in urban planning and design, ranging from theoretical discussions, methodological innovation and empirical investigation.
Following the call for submissions for this special issue, several paper sessions were organized at the 2017 International Conference on Geoinformatics as well as at the annual meetings of the American Association of Geographers. All papers have undergone a rigid abstract selection from about 60 submissions and a follow-up peer review process. This special issue includes contributions from seven articles with very different perspectives by academics working in architecture, geography, sociology, urban planning, criminology, public policy and economics. The authors were encouraged to stress the planning and practical implications of their results with the hope of moving towards ‘use-inspired’ basic research (Stokes, 1997) as well as bridging research and practice. We have identified three interrelated themes from the work published in this special issue: (1) social networks in space, (2) joint socio-spatial effects and (3) integrated spatial and network analytics.
Integrating social networks and spatial analyses
Social networks in space
This body of literature is concerned with how space matters for social interactions. While this literature has grown rapidly in both scope and depth, recent efforts have been made to explicitly link social interactions with morphological properties. We see these morphological properties as potential ‘bridges’ between research and design communities. In this special issue, Boessen et al. (2018) relate ‘three D’s’ in urban and transport planning – density, diversity and design – to the number and location of inter-personal networks. They take a multi-scale and multi-dimensional approach, exploring how the built environment affects three different types of social networks (i.e. socializing friendship ties, core ties for important matters, as well as kin ties) at different geographical scales. The analysis applies a series of regression models based on a large-scale egocentric network in the Western USA.
The papers by Zhao and Wang (2018) as well as Zhou et al. (2018) are concerned with (re)conceptualizing planning outcomes and goals using network terms. More specifically, Zhao and Wang (2018) introduce a network-explicit approach to characterizing the degree of segregation in a Chinese community. Their analysis relies on a two-mode social network, where nodes are individual people (both local and migrant residents) and routine venues. The results highlight a modest level of social interactions between local and migrant residents as well as how socially and spatially different segregation is in the Chinese and North American contexts. Furthermore, Zhou et al. (2018) have employed emerging mobile phone big data to explore the jobs-housing balance and employment self-containment, which are central concerns of land use planning.
Joint socio-spatial effects
This strand of analysis explores how the built form and society interact with one another, which influences other aspects of urban performance. As Andris (2016: 2009) put it, ‘we are simultaneously born into a geographic landscape and a social network (SN) … Throughout our lives, we use the intertwined, inextricable systems of the SN and the geographic landscape to grow and develop’. The papers by Kim et al. (2018) and Gibbons et al. (2018) in this special issue are concerned with how spatial and social factors jointly affect our urban life. More specifically, Kim et al. (2018) argue that individual travel behaviours are affected by both intra-household interactions between family members as well as inter-household interactions between nearby households, which are, in turn, conditioned by the built environment and socioeconomic status. The analysis is based on a detailed travel survey in the USA and develops a novel discrete choice model to link travel activities in different transport modes with individual, household as well as neighbourhood characteristics. One of the key findings is that social interactions and the built environment jointly affect the willingness to use active transport modes (e.g. walking and cycling). Another example is the paper by Gibbons et al. (2018), in which the authors explore how the effects of gentrification and social media networks are intertwined within individual communities.
Integrated spatial and network analytics
On the analytical front, attempts have been made to synthesize networks and spatiotemporal methods to capture both geographical and social dimension of urban dynamics. The paper by Li et al. (2018) proposes a multi-scale (individual, local, meso, and global) and multi-faceted (space, time, and network) framework for analysing urban trajectory dynamics. The usefulness of this framework is demonstrated with two large human trajectory data sets that are derived from GPS tracks and mobile phone records. While Li et al. (2018) aim to synthesize spatial and network analytics into one framework, another approach is adopted by Sevtsuk and Kalvo (2016), who integrate network-based measures into conventional spatial analysis. More specifically, they revise the conventional Huff model in retail planning and replace straight-line distances with measures based on street networks. The revised model is used to compare planning scenarios for new town retail centres in Singapore.
Future studies
The possible avenues for future studies presented in this special issue echo those identified from related fields (Adams et al., 2012; Andris, 2016). Noting that papers in this issue do not cover the full scale of spatial social network urban research, we have argued elsewhere that the following research avenues are also noteworthy (Ye and Liu, 2018). First, there is a need to develop a theoretical framework to integrate quantitative and qualitative analyses. While quantitative models such as those demonstrated in this special issue can characterize the overall patterns of and relationships between physical and social spaces, they do not necessarily inform us about individual agency and practice, the multi-faceted nature of social connections, and the richness and complexity of social–spatial interactions. Therefore, conventional qualitative approaches such as interviews, sketching and site visits are still useful in providing nuanced understandings of how these dynamics unfold in specific contexts (Ye et al., 2017). Still, GPS-enabled videos, pictures, and geo-narratives at the fine scale are transformative field approaches for retrieving contextual information for understanding human interaction and networking within the built environment (Curtis et al., 2016). In addition, we need to cautiously draw analytical conclusions given the bias and uncertainty of various spatial social network data types (Shaw et al., 2016). Secondly, more attention should be paid to how the nodes and links are defined in networks because different definitions may lead to diverse network patterns even within the same data set (Liu and Derudder, 2013) and that theoretically grounded definitions of nodes and links are needed to circumvent the network version of the Modifiable Area Unit Problem (van Meeteren and Poorthuis, 2017).
Third, data-challenged built environments warrant notice because ignoring the co-existence of data-rich communities and data-poor communities in a city would lead to many distorted analytical results that are possibly meaningless in policy implementation and program evaluation. For example, a number of major urban centres have depressed communities with a large percentage of low-income population with minimal technical skills or lacking reliable access to the Internet, leading to a scarcity in retrieved social data.
Fourth, as for the more technical side, the integration of spatial and social network analyses requires proper software environments. On the one hand, while spatial design tools such as Space Syntax and sDNA have long been used in design and planning practices, they are more aptly applied to technological and infrastructure networks. Moreover, Space Syntax focuses on street networks and/or architectural plans as it argues that urban form and activities overlap (Hillier and Hanson, 1984). While this holds true for ‘old cities', future studies may want to explore whether there is a form-function separation in the current digital society. On the other hand, common geographic information system tools are less well-prepared for dealing with network data and are often designed for special purposes such as the identification of trading areas and the planning of bus routes.
Domain users and decision makers need to store, visualize and compute big and dynamic spatial social network data in the context of a real-time city. Several papers in this issue are already experimenting with such data. However, the data scale and dimensions impose severe technical challenges for them, as these might lack hard and soft infrastructures supporting these data-driven approaches (Liu and Long, 2016; Ye et al., 2016). In addition, the raw spatial social network data in the built environment are often big, heterogeneous, and unstructured with noise. It can therefore be challenging to process such data and apply the findings to policy making. There are urgent needs to develop robust and effective methods of analyzing such data to enhance our understanding of dynamic human activities, relationships, and their changes over time. (Shaw et al., 2016: 1689).
Hence, the implementation of public-licensed big data analytics software with free access would facilitate the utilization of social interaction data at an unparalleled spatial and temporal granularity level towards more policy-relevant human-centred urban science. The open data and open source toolkit movement can advance such research activities and build an interdisciplinary community of worldwide urban researchers and educators related to spatial social network studies (Ye, 2018).
Footnotes
Acknowledgements
This material is partially based upon work supported by the National Science Foundation under Grant Nos. 1416509, 1535031, 1637242, and 1739491. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.