Spatial Behavior, Scaling Laws and Built Environment Data

Fine-scale, geocoded neighborhood socioeconomic status (nSES) data are foundational inputs for urban studies and spatial inequality research, yet remain scarce in emerging economies that together account for over 70% of the global population. Here, we use publicly available satellite and street view imagery to estimate nSES at fine spatial resolution, proposing an ensemble regression framework that unites multi-view images to measure intra-urban inequality. We estimate eight indicators—spanning household registration (hukou), housing, income, employment, and education—as numeric values rather than broad quantile bands, enabling direct cross-neighborhood comparison. Family and individual SES from a representative survey serve as training data. The proposed fusion model outperforms single ensemble regression models and single-view image-based models in mean absolute percentage error. To interpret image- based prediction, we apply SHAP-based feature attribution and modality-contribution analysis, identifying the built- environment features that drive predictions and revealing that satellite and street view imagery play complementary rather than redundant roles. Neighborhood inequalities at traffic analysis zone and township scales in central Beijing are visualized for the first time. We further transfer the trained model to Shanghai, where no comparable public nSES dataset exists, providing a limited proof-of-concept against real-estate data for two of the eight indicators and outlining a replicable pathway for data-scarce cities.

Buildings serve as the basic cells of a city, forming a complex urban system that integrates physical and social characteristics. While urban analytics and simulation have traditionally been constrained by data availability and computational complexity , recent advancements in sensing, AI and high performance computing have made building-scale inquiry increasingly feasible. This perspective proposes and highlights Building-Scale Urban An alytics and Simulation through four key pillars. First, it calls for a unified modeling standard that aligns the semantic depth of building engineering with the geographic breadth of urban science. Se cond, it highlights the need for constructing comprehensive building datasets with full spatiotemporal coverage through AI-driven data foundations. Third, it advocates for the use of generative AI to diagnose building quality and efficacy. Fourth, it proposes a bottom-up simulatio n paradigm that integrates AI with generative rules to model complex morphological and behavioral transformations. This vision provides a critical bridge between building studies and urban studies.

Urban heat adaptation strategies rely on assumptions about how people use public spaces, yet actual behavioral responses remain largely unexplored. Using 360 ◦ video tracking and AI-powered object detection (YOLOv10) across 11 Beijing parks, we reveal counter-intuitive patterns that challenge conventional cooling design. While rising temperatures suppress overall park use, specific facilities exhibit unexpected effects: sports courts sustain and even increase youth activity under heat stress, suggesting functional needs override thermal discomfort. Conversely, sheltering elements prove essential for vulnerable populations, with pavilions and shaded seating promoting critical static rest behaviors. Our multi-scale analysis demonstrates that macro-environmental context (e.g., the greenery level of the park ’ s surroundings) modulates these micro-scale effects (e.g., the influence of specific in-park facilities), with parks in less vegetated surroundings becoming vital cooling refuges. These behavioral insights enable a fundamental shift from reactive cooling strategies to proactive design that strate - gically leverages human adaptation patterns, providing empirical foundations for climate-resilient urban spaces.

1Scientific Data | (2025) 12:430 | https://doi.org/10.1038/s41597-025-04730-5 www.nature.com/scientificdata CMaB: a Multi-attribute Building Dataset of China Yecheng Zhang 1,3, Huimin Zhao1,3 & Ying Long 1,2 ✉ Rapidly acquiring three-dimensional (3D) building data, including geometric attributes like rooftop, height and orientations, as well as indicative attributes like function, quality, and age, is essential for accurate urban analysis, simulations, and policy updates. Current building datasets suffer from incomplete coverage of building multi-attributes. This paper presents the first national-scale Multi- Attribute Building dataset (CMAB) with artificial intelligence, covering 3,667 spatial cities, 31 million buildings, and 23.6 billion m² of rooftops with an F1-Score of 89.93% in OCRNet-based extraction, totaling 363 billion m³ of building stock. We trained bootstrap aggregated XGBoost models with city administrative classifications, incorporating morphology, location, and function features. Using multi- source data, including billions of remote sensing images and 60 million street view images (SVIs), we generated rooftop, height, structure, function, style, age, and quality attributes for each building with machine learning and large multimodal models. Accuracy was validated through model benchmarks, existing similar products, and manual SVI validation, mostly above 80%. Our dataset and results are crucial for global SDGs and urban planning.

The Fourth Industrial Revolution has been profoundly reshaping economics, socie- ties, and individuals, and transforming cities through the integration of a range of emerging technologies, represented by mobile internet, big data, robotics, artificial intelligence (AI), immersive media (virtual reality/augmented reality/mixed reality), and the Internet of Things (IoT) (Moffitt 2018; Schwab 2017). The fusion and interac- tion of these technologies are not only playing an increasingly essential role in reshap- ing urban daily life and space but also creating unprecedented opportunities to accurately depict and understand cities, thereby further supporting more refined-scale urban planning and management practice (Long and Zhang 2021; Engin et al. 2020). The state-of-the-art studies examining how technology shapes urban develop- ment revolve around three key aspects. Some scholars acknowledge the merits of big data derived from Information and Communication Technology (ICT) in pro - viding unprecedented opportunities to sense fine-scale socioeconomic activity and mobility (Liu et al. 2015), explore the new science of cities (Batty 2013a), and enhance decision making and management (Karimi et al. 2021).

In recent years, pandemics have become one of the most significant challenges due to their huge socio-economic impacts. Fortunately, smart technologies have provided new ideas to fight against them. Many studies have focused on analyzing particular technologies applied in pandemics, but few have systematically discussed the difference and the relationship among multiple perspectives. China is well represented in the development of technologies and pandemic responses. Therefore, this paper uses China’s response to COVID-19 as an empirical study to systematically review the application of smart technologies and build a case base from multiple perspectives. A total of 1,102 cases from 14 technologies were collected from January 2020 to June 2020 after screening, and a series of analyses were conducted in terms of types, scales, stages, and targets. The result shows various subjects participated in pandemic responses using smart technologies. General technologies such as Big Data and Mobile Internet are most widely used. Besides, most technologies are used on the country or district/city scales and focus on the prevention and control of pandemics. There are significant differences in the penetration of technologies among different perspectives. We hope to provide a reference for applying smart technologies against pandemics in the future.

The pandemic of physical inactivity has become a global issue. The WHO recommends engag- ing in physical activity (PA) as an important, modifiable behavior for preventing noncommu- nicable chronic diseases, such as obesity, diabetes, and cardiovascular disease, and for promoting public health. While many studies have explored ways to encourage PA through education [1], social support [2], and health promotion strategies [3], an increasing number of studies in both urban planning and public health have begun to examine the socioecological context of PA, suggesting that the environment in which PA takes place should be considered. The accumulation of evidence demonstrates that people who live in walkable neighborhoods, which are characterized by higher population density, a mix of different urban functions, interconnected street networks, and access to shops and services, public transport, parks and recreational facilities, tend to be more physically active [4–6]. When considering the factors PLOS ONE PLOS ONE | https://doi.or g/10.137 1/journal.po ne.02605 70 December 31, 2021 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Chen L, Zhang Z, Long Y (2021) Association between leisure-time physical activity and the built environme nt in China: Empirica l evidence from an acceleromete r and GPS-based fitness app. PLoS ONE 16(12): e0260570.

Zones, cells, and parcels have long been regarded as the main units of analysis in urban modeling. However, only limited attention has been paid to street-level urban modeling. The emergence of fine-scale open and new data available from various sources has created substantial opportunities for research on urban modeling at the street level, particularly for modeling the spatiotemporal process of urban phenomena. In this paper, the street is adopted as the spatial unit of an urban model, and a conceptual framework for such modeling based on cellular automata is proposed. The validity of the proposed framework is verified by an empirical application to the urban space within the Fifth Ring Road in Beijing from 2014 to 2018. The results show that the density of points of interest simulated by the cellular automata model for 2018 is basically consistent with the actual distribution according to direct observation, and there is no significant difference in the Corresponding author: Ying Long, School of Architecture and Hang Lung Center for Real Estate, Tsinghua University; Key Laboratory of Eco Planning & Green Building, Ministry of Education, China. Email: ylong@tsinghua.edu.cn EPB: Urban Analytics and City Science 2020, Vol. 47(8) 1418–1439 ! The Author (s) 2020 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/2399808320942777 journals.sagepub.com/home/epb proportion of high, medium, and low points of interest density streets between different ring roads. In addition, the deviation rate and Kappa index are 0.1171 and 0.97, respectively, indicating the proposed model can replicate historical patterns well and predict the transition of points of interest density at the street level. Subsequently, we considered three scenarios, adopting 2018 as the base year and using the proposed model to simulate the distribution of points of interest density in 2022 and the changes in points of interest density from 2018 to 2022. The conceptual framework and empirical application also provide support for urban planning and design based on the integration of linear public space and big data.

Existing studies have extensively used spatiotemporal data to discover the mobility patterns of various types of travellers. Smart card data (SCD) collected by the automated fare collection systems can reflect a general view of the mobility pattern of public transit riders. Mobility patterns of transit riders are temporally and spatially dynamic, and therefore difficult to measure. However, few existing studies measure both the mobility and stability of transit riders’ travel patterns over a long period of time. To analyse the long-term changes of transit riders’ travel behaviour, the authors define a metric for measuring the similarity between SCD, in this study. Also an improved density-based clustering algorithm, simplified smoothed ordering points to identify the clustering structure (SS-OPTICS), to identify transit rider clusters is proposed. Compared to the original OPTICS, SS-OPTICS needs fewer parameters and has better generalisation ability. Further, the generated clusters are categorised according to their features of regularity and occasionality. Based on the generated clusters and categories, fine- and coarse- grained travel pattern transitions of transit riders over four years from 2010 to 2014 are measured. By combining socioeconomic data of Beijing in the year of 2010 and 2014, the interdependence between stability and mobility of transit riders’ travel behaviour is also discussed.

Physica A 525 (2019) 1429–1439 Contents lists available at ScienceDirect Physica A journal homepage: www.elsevier.com/locate/physa Reinvestigating China’s urbanization through the lens of allometric scaling Wei Lang a,b,c,1, Ying Long d,1, Tingting Chen a,b,c,1, Xun Li a,b,c,∗ a Department of Urban and Regional Planning, School of Geography and Planning, Sun Yat-sen University, Guangzhou, 510275, China b Urbanization Institute, Sun Yat-sen University, Guangzhou, 510275, China c China Regional Coordinated Development and Rural Construction Institute, Sun Yat-sen University, Guangzhou, 510275, China d School of Architecture and Hang Lung Center for Real Estate, Tsinghua University, Beijing 100084, China h i g h l i g h t s • Urbanization in China follows allometric scaling law. • Scaling exponent varies at the different stages neither static state nor isometric. • Urban system allometric scaling evolves to align with the theoretical assumption. • Government-led developments deviates the scaling relations for a short term. • Driving force of complex system shifts leading role at various urbanization stages. a r t i c l e i n f o Article history: Received 25 October 2018 Received in revised form 3 February 2019 Available online 8 April 2019 Keywords: Urbanization Allometric scaling City size Urban system China a b s t r a c t Cities are complex systems;

Urbanization is a complex spatial phenomenon involving signi ficant compositions and interactions in land use. Yet, only few studies have quantitatively examined multidimensional urban land-use patterns with insights into land use policy, particularly in the context of China’ s rapid urbanization process. This paper aims to investigate the urban land-use patterns in China by employing multiple measurements with multi-sourced data, including Spatial Entropy and Dissimilarity Index, and a combination of cellular-automata (CA) modeling and Structural Equation Modeling (SEM). The results show that land-use patterns in China are characterized from more mixed (Beijing, Shanghai) to less segregated (Xiangyang, Tangshan, and Guiyang), and the most segregated (Chongqing), which can be categorized into three typical types: economically led, government led, and geo- graphically constrained. The findings also indicate that residential sector has correlation with GDP and urban built-up area; public sector is driven by GDP, urban built-up area, and paved road area; and commercial sector is related to GDP and paved road area. Furthermore, land-use patterns are not only determined by economic forces, but also subject to China ’s land policies that formulated based on its unique social and political characteristics. It reveals the complex spatial characterization of urbanization in China, where government still plays an important role in facilitating the land use allocation. The research sheds light on understanding land use policy for land-use patterns recon figuration in the context of New-Type Urbanization towards better planning and governance.

Most analyses of travel patterns are based on the assumption of isolated individuals and ignore interpersonal relationships between travelers. In this paper, we develop a straightforward method to identify group travel behavior (GTB), de fined as two or more persons intentionally traveling together from a single origin to a single destination, with public transit smart card data based on proxemics theory. We apply our method to Beijing to reveal the patterns of GTB, using all records generated by the subway system during a one-week period in 2010. Our data and method do not allow a reliable estimate of GTB share in overall travel, but do enable a description of the characteristics and the spatiotemporal pattern of GTB. The results reveal that the group size and GTB frequency follow a long tail distribution: far more people travel in small groups than in large groups and far more group travelers can be observed carrying out only one group trip than travelers making multiple group trips. Group trips tend to occur in weekends, in afternoons, and during public holidays. Furthermore, stations and lines serving leisure destinations show the highest GTB scores. We conclude that the GTB pattern is distinctly di fferent from the pattern of individual travel in terms of both time and space, and is essentially in fluenced by urban land uses surrounding subway stations.

Extreme conditions often capture our attention and point to important underlying mechanism; we have learned a great deal about our cities by examining the extremes, such as the emergence and dynamics of the most dominant city of a nation, the most depressed city of a region, as well as the most popular gateway city among immigrants. In the past decade or so, against the backdrops of the global financial crisis, increased numbers of the unemployed, self-employed and part-time workers, the rise of telecommuters as well as the relocation of low-paying jobs, extreme commuters have received increasing academic and public attention in recent years. As extreme commuting accounts for an increased portion of daily residential trips, recent analysis starts to look into travelers making unusually long, early, late, and/or frequent trips, which have discretely explored or described by Barr, Fraszczyk, and Mulley (2010); Gregor (2013); Jones (2012); Landsman (2013); Marion and Horner (2007); Moss and Qing (2012); Rapino and Fields (2013); U.S. Census (2005) . As scholarly work on extreme travelers has been largely devel- oped based on North American and European cities, we are inter- ested in extending the framework to understand extreme trips in China.

Existing studies have extensively used temporal-spatial data to mining the mobility patterns of different kinds of travel- ers. Smart Card Data (SCD) collected by the Automated Fare Collection (AFC) systems can reflect a general view of the mobility pattern of the whole bus and metro riders in urban area. Since the mobility and stability are temporally and spatially dynamic and therefore difficult to measure, few work focuses on the transition of their travel pattern between a long time interval. In this paper, an overview of the relation between stability and regularity of public transit riders based on SCD of Beijing is presented first. To ana- lyze the temporal travel pattern of urban residents, travelers are classified into two categories, extreme and non-extreme travelers. We have two lines for profiling all cardholders, rule based approach for extreme and improved density-based clustering method for non-extreme. Similar clusters are ag- gregated according their features of regularity and occasion- ality. By combining transition matrix of passenger’s tempo- ral travel pattern and socioeconomic data of Beijing in the year of 2010 and 2014, several analyses about resident’s tem- poral mobility and stability are presented to shed lights on the interdependence between stability and mobility in the time dimension. The results indicate that passengers’ reg- ularity is hard to predict, extreme travel patterns are more vulnerable and overall non-extreme travel patterns nearly stay the same. Categories and Subject Descriptors I.5 [Pattern Recognition]: Clustering; H.2.8 [ Database Management]: Database Applications— Data mining

Characterizing Evolution of Extreme Public Transit Behavior Using Smart Card Data Zhiyong Cui Department of Civil and Environmental Engineering Univesity of Washington Seattle, W A, US Email: zhiyongc@uw.edu Ying Long Beijing Institute of City Planning Beijing, China Email: longying1980@gmail.com Ruimin Ke Department of Civil and Environmental Engineering Univesity of Washington Seattle, W A, US Email: ker27@uw.edu Yinhai Wang Department of Civil and Environmental Engineering Univesity of Washington Seattle, W A, US Email: yinhai@uw.edu Abstract—Existing studies have extensively used temporal- spatial data to mine the mobility patterns of different kinds of travelers. Smart Card Data (SCD) collected by the Automated Fare Collection (AFC) systems can reflect a general view of the mobility pattern of the whole bus and metro riders in urban area. Most existing work focusing on mobility pattern usually ignore a special group of people who travel in abnormal patterns or mechanisms. In this paper, we focus on the evolution extreme transit behaviors of travelers in urban area by using SCD in 2010 and 2014. We have several aspects of descriptive statistics of the SCD with a view to better understanding the dynamic process and evolution of the extreme transit behavior.

Land-use pattern is one of the key issues in the compilation of urban master plans. In China, government, planners, and residents, all with various requirements and preferences, are the main agents participating in this process. Among them, planners play a role in negotiating with related agents and then establishing land-use patterns. In this paper we propose a planner agent framework to support land-use pattern scenario analysis (LUPSA), based on existing planning support system (PSS) research. Planner agents are divided into three types: nonspatial planner agent (NP A), spatial planner agent (SP A), and chief planner agent (CP A). The NP A is responsible for formulating special plans (such as transport, municipal public facilities, or nature reserve plans) that correspond to available data (such as road network, public facilities, and nature reserve patterns) from LUPSA. The SP A is responsible for establishing land-use patterns. The SP A considers constraints of local development conditions and communicates and coordinates with the NP As to confirm formulated special plans that can support the implementation of the established land-use pattern. The CP A is responsible for negotiating with the government agent to ensure the reasonability of comprehensive constraints, establishing the final land-use pattern based on an evaluation of established scenarios by several SP As, then determining it after a public participation process involving local residents. We initially tested this framework in a hypothetical city, then did an experiment in Beijing. Results show that the proposed planner agent framework is suitable for LUPSA.

As an important part of la nd use/cover change (LUCC), historical LUCC in long time series attracts much more attention from scholars. Currently, based on the view of combining the overall control of cropland area and ′top-down′ decision-making behav- iors, here are two global historical land- use datasets, generally referred as the Su stainability and the Gl obal Environment dat asets (SAGE datasets) and History Database of the Global Environment datasets (HYDE datasets ). However, at the regional level, these global datasets have coarse resolutions and inevitable errors. Considering various factors that influenced cropland distribution, including cropland connectivity and the limitation of natural and human factors, this study developed a reconstruction model of historical cropland based on constrained Cellular Automaton (CA) of ′bottom-up′. Then, an available labor force index is used as a proxy for the amount of cropland to inspect and calibrate these spa tial patterns. Applied the reconstruction model to Shandong Province, we reconstruct ed its spatial distribution of cropland during 8 periods. The reconstructe d results show that: 1) it is properly suitable for constrai ned CA to simulate and reconstruct the spatial distri bution of cropland in traditional cultivated region of China; 2) compared with ′SAGE datasets′ and ′HYDE datasets′, this study have formed higher-resolu tion Boolean spatial distribution datasets of historical cropland with a more definitive concept of spatial pattern in terms of fractional format.

Research on reconstructing spatial distribution of historical cropland over 300 years in traditional cultivated regions of China Yang Xuhong a, Jin Xiaobin a,b,⁎,G u oB e i b e ia, Long Ying b,c,Z h o uY i n k a n ga,b a School of Geographic and Oceanographic Sciences, Nanjing University, Nanjing 210023, China b Natural Resources Research Center of Nanjing University, Nanjing 210023, China c Beijing Institute of City Planning, Beijing 100045, China abstractarticle info Article history: Received 18 June 2014 Received in revised form 19 January 2015 Accepted 17 February 2015 Available online 24 February 2015 Keywords: China traditional cultivated region historical cultivated land spatial distribution reconstruction LUCC Constructing a spatially explicit time series of historicalcultivated land is of upmost importance for climatic and eco- logical studies that make use of Land Use and Cover Change (LUCC) data. Some scholars have made efforts to sim- ulate and reconstruct the quantitative information on historical land use at the global or regional level based on “top–down”decision-making behaviors to match overall cropland area to land parcels using land arability and uni- versal parameters. Considering the concentrated distribution of cultivated land and various factors influencing cropland distribution, including environmental and human factors, this study developed a“bottom–up”model of historical cropland based on constrained Cellular Automaton (CA).

Environment and Planning A 2014, volume 46, pages 2051 – 2054 doi:10.1068/a140031g Featured graphic. Visualizing the minimum solution of the transportation problem of linear programming (TPLP) for Beijing’s bus commuters A number of scholars (see Horner, 2002; White, 1988) have established a framework for analyzing the efficiency of regional commuting patterns. Typically, this framework has a minimum and maximum commute (T min and T max). Commuting is considered excessive if actual commuting (Tact) deviates from Tmin in a given city region. Tmin assumes that individuals commute, on average, to the closest possible workplace in terms of some measure of zonal separation (eg, distance, time) while T max assumes the opposite. However, while many scholars have calculated Tmin and Tmax for various cities, none has attempted to map their solutions. Thus, we still do not have any cartographic knowledge of how the geography of flows associated with T min compares with Tact. Using one week of bus users’ origin and destination (OD) flow data attained from Smartcards in Beijing for 2008, we not only calculated T act and Tmin for bus commuters in the city but also mapped the corresponding commuting flows. The figures below are our mapping results. For technical details regarding how we derive and validate OD flow data for bus commuters in Beijing, please refer to Long et al (2012), while Horner (2002) can be consulted for details on how to calculate T act, Tmin, and Tmax.