The Evolution of the Interaction Between Urban Rail Transit and Land Use: A CiteSpace-Based Knowledge Mapping Approach

Abstract

Urban rail transit is a key enabler for optimizing urban spatial structures, and its interactive relationship with land use has long been a focus of attention. However, existing studies suffer from scattered methodologies, a lack of systematic analysis, and insufficient dynamic insights into global trends. This study comprehensively employs CiteSpace, VOSviewer, and Scimago Graphica to conduct bibliometric and knowledge map analysis on 1894 articles from the Web of Science database between 2004 and 2024, focusing on global research trends, collaboration networks, thematic evolution, and methodological advancements. Key findings include the following: (1) research on rail transit and land use has been steadily increasing, with a significant “US-China dual-core” distribution, where most studies are concentrated in the United States and China, with higher research density in Asia; (2) domestic and international research has primarily focused on themes such as the built environment, value capture, and public transportation, with a recent shift toward artificial intelligence and smart city technology applications; (3) research methods have evolved from foundational 3S technologies (GIS, GPS, RS) to spatial modeling tools (e.g., LUTI model, node-place model), and the current emergence of AI-driven analysis (e.g., machine learning, deep learning, digital twins). The study identifies three future research directions—technology integration, data governance, and institutional innovation—which provide guidance for the coordinated planning of transportation and land use in future smart city development.

1. Introduction

As the pace of global urbanization accelerates, urban rail transit (URT) plays a key role as an efficient and sustainable mode of transportation in optimizing urban spatial structures, alleviating traffic congestion, and promoting low-carbon development [1,2]. At the same time, the dynamic evolution of land use patterns directly impacts passenger distribution, operational efficiency, and socio-economic benefits of rail transit [3,4,5]. The interaction between the two has become a hot topic of research in fields such as urban planning, transportation engineering, and land management [6,7,8].

Despite a growing body of literature exploring this interaction, several key challenges persist in the current research landscape.

First, although early classical literature reviews have laid a systematic foundation for the disciplinary framework [9,10], the rapid evolution of transportation technology and urban spatial forms in recent years has made it imperative for the academic community to conduct a phased integration and reassessment of research findings and cutting-edge trends [11,12]. Current research still exhibits significant disciplinary fragmentation, with most studies focusing on specific regional cases (e.g., North American polycentric urban systems) [10] or micro-level issues (e.g., TOD [13], algorithm optimization of LUTI models [14], environmental assessment [15]), lacking an integrated analytical framework capable of synthesizing technological evolution logic, and spatial interaction effects [16,17].

Second, traditional literature reviews primarily rely on qualitative analysis, which limits their repeatability and objectivity when tracking the evolution of research topics, co-authorship networks, and methodological advancements [8,18]. As a result, they struggle to conduct a systematic assessment of the global evolution of core theoretical paradigms. In recent years, with the innovation of bibliometric tools, the academic community has gradually turned to using visualization analysis software to systematically mine knowledge from multiple sources of academic data [19,20]. Unlike existing studies that focus on statistical integration of general data, this paper focuses on the SCI-EXPANDED and SSCI journal databases and uses structured bibliometric methods to improve the original limited precision problem, thereby achieving an in-depth description of the cutting edge of this field and accurate tracking of theoretical developments [21].

Third, although emerging technologies such as AI, big data analysis, and digital twins have provided new avenues for URT-land use research, existing reviews suffer from timeliness lag [22]. Therefore, this study systematically reviews the latest advances in intelligent technologies such as machine learning [23] and deep learning [24] under the theoretical framework of coordinated development of rail transit and land use. Based on the existing literature review [25], this study provides a new theoretical foundation for further understanding the multi-level relationship between this field and new technologies.

This study conducts a comprehensive bibliometric and knowledge mapping analysis of 1894 academic publications from the Web of Science (WOS) Core Collection spanning 2004–2024. Leveraging CiteSpace, VOSviewer, and Scimago Graphica, the study investigates spatio-temporal publication trends, author collaboration networks, thematic clusters, and technical method adoption. The guiding research questions are as follows:

• RQ1: What are the global research trends in this field?
• RQ2: How have key research themes evolved over time?
• RQ3: How are emerging technologies changing research methods in this field?

This study makes significant contributions to the field of urban rail transit and land use interaction by conducting a systematic literature review and focusing on key research questions. Research reveals the dual-core spatial publication structure of China and the United States, clearly showing the distribution of global research forces. It identifies the evolving research theme trend from TOD construction to value capture and travel behavior, and reviews the development trends of methods such as machine learning and deep learning. These findings provide scholars with references for understanding research trends and innovating research methods, and integrate diverse content to establish a comprehensive research perspective. This offers urban planners and policymakers highly practical guidance for developing sustainable urban transport policies.

This study is structured to address the three core research questions through progressive analysis (Figure 1). Section 2 details the data and bibliometric methods. Section 3 responds to RQ1 and RQ2 by mapping global research trends. Section 4 further answers RQ2 and RQ3 through an interpretation of thematic evolution and technological progress based on keyword clustering. Each subsection corresponds to major clusters—built environment, value capture, travel behavior, and planning technologies—revealing their interconnections. Section 5 synthesizes insights and proposes future directions grounded in the analytical findings.

2. Materials and Methods

2.1. Data Collection

To examine the evolution of knowledge regarding the relationship between urban rail transit and land use, we compiled the pertinent literature from the Science Citation Index Expanded (SCI-EXPANDED) and Social Sciences Citation Index, both of which are integral databases within the Web of Science Core Collection. The literature search was conducted on 26 November 2024, using a targeted search method focused on the following keywords: (urban rail transit OR metro systems OR subway systems OR light rail transit OR urban public transportation) AND (land use OR land utilization OR land planning), with a time restriction from 1 January 2004, to 26 November 2024, to search the database content. Secondly, the initially obtained database literature was cleaned and deduplicated, resulting in 1894 articles. As an important component of the urban transportation system, urban rail transit systems play a crucial role in connecting various urban areas, while the evolution of land use patterns reflects the dynamic development of urban space. This study employs CiteSpace knowledge mapping technology to explore the research trajectory of the interaction between urban rail transit and land use.

2.2. Bibliometric Methods

The advent of the era of big data has placed researchers in a vast sea of literature. However, the continuous advancement of information technology and the growing capabilities of computer processing have opened new avenues for the refined and visual analysis of literature data. At present, more than 10 types of software are available for mapping foreign literature, each offering distinct advantages. For example, VOSviewer excels in topic clustering, providing clear and detailed visualizations, while HistCite organizes literature citations in a time-based network format. CiteSpace, renowned for its powerful co-citation analysis, has become a key tool in over 60 fields, including computer science, information science, and medicine, due to the ongoing refinement of its algorithms and functions. This study employed both CiteSpace 6.3.R1 and VOSviewer 1.6.20. to conduct a visual analysis of articles in the field, extracting cutting-edge research hotspots, connections, and trends. However, while CiteSpace is effective in mechanically calculating connections between articles and identifying prominent keywords and research topics, the specifics of the literature require manual reading and analysis for deeper understanding. In this study, CiteSpace 6.3.R1 was used to eliminate duplicates from the dataset, after which the author and keyword networks were explored. Additionally, Scimago Graphica 1.0.50.0 was utilized to visualize the national and regional distribution of research activity.

3. Performance Analysis

3.1. Spatial-Temporal and Source Distribution Characteristics

This study provides an overview of the global evolution and research landscape of the literature over the past two decades. It analyses the spatio-temporal dynamics of academic output (RQ1), including publication trends, national and institutional distribution patterns, and contributions from core journals. This reveals the key characteristics of the research field. This reveals the geographical and institutional diffusion pathways of related research. This lays the groundwork for subsequent studies on co-cited literature (RQ2 and RQ3).

3.1.1. Publication Trends

Figure 2 shows that the development of this research field can be illustrated by analyzing the number of papers published in recent years. The development of rail transit and land use research from 2004 to 2024 reflects distinct phases shaped by urbanization trends and policy shifts. There has been an overall upward trajectory in annual publications, rising from six to 209 papers.

During the initial slow growth phase (2004–2011), limited data acquisition methods and an insufficient research foundation resulted in low awareness of land planning, particularly surrounding urban rail transit stations (the Transit-Oriented Development (TOD) concept had not yet been widely adopted), leading to limited related research. The subsequent rapid growth phase (2011–2018), driven by accelerated urbanization, particularly in East Asia, saw a dramatic increase in research activity, with studies examining themes such as destination accessibility and transport network proximity, and the global promotion of TOD models [10]. China’s 2014 ‘New Urbanisation’ strategy further stimulated this expansion by emphasizing coordinated rail–city development. The current high-quality growth phase (2018 to the present day) marks a strategic shift from quantity to sophistication. This is influenced by stricter Chinese regulations, such as the 2018 National Development and Reform Commission Document No. 52, which imposed financial safeguards and redirected the focus from construction scale to operational efficiency. Meanwhile, technological advancements have introduced big data analytics, machine learning and smart city concepts into research methodologies. This evolution clearly aligns research trends with policy priorities, progressing from initial technical explorations and mid-phase planning analyses to the current focus on smart, sustainable development.

3.1.2. Analysis of Countries

This study covers a wide range of the literature involving 56 countries across seven continents. As Figure 3 shows, the research results exhibit significant spatial distribution characteristics: they are concentrated in the Northern Hemisphere, primarily in countries with relatively favorable economic conditions, forming a dual-core structure centered on China and the United States. China and the United States stand out among these countries, with the highest research output: 737 and 626 papers were published by these two countries, respectively, accounting for 39% and 33% of the total literature. Canada and Australia follow with 134 and 126 papers, respectively. Although European countries have relatively few publications, they demonstrate strong interconnections, indicating high levels of academic collaboration within the region. Despite its small land area, Singapore has established extensive research collaboration networks with China, Indonesia, Malaysia and other countries thanks to its geographical location. Australia and New Zealand, despite their geographical distance from other continents, have continued to innovate in urban rail transit and land use research, maintaining close academic ties with European and American countries. By analyzing the spatial distribution patterns of literature and transnational cooperation networks around the world, this study reveals differences in academic activity levels and cooperation models between different regions.

3.1.3. Institutional Analysis

Figure 4 shows the institutions with the highest number of new publications in the field of rail transit and land use from 2004 to 2024. The institution that has contributed the most to research in this field is Tongji University, which ranks first with 54 articles. Second is Wuhan University, with 52 papers published, followed closely by the University of Hong Kong and Southeast University, which have published 46 and 45 articles, respectively. In addition, Beijing Jiaotong University, with its advantages in transportation, has published 40 articles. Other institutions around the world, such as the University of Toronto, MIT, the University of California, Berkeley and the University of North Carolina, have continued to conduct interdisciplinary research in related fields, providing case studies for related research in developed countries around the world. Overall, research in this field is dominated by the US and China, which is consistent with the ‘US-China dual-centre’ pattern of global research output shown in Figure 3.

3.2. Collaborative and Intellectual Networks

Building upon the macro-level spatio-temporal and institutional insights presented in Section 3.1, Section 3.2 delves deeper into the collaborative and intellectual structures that shape the field of urban rail transit and land use integration. The analysis is conducted from several dimensions, including highly co-cited documents and co-citation author networks. This allows us to trace the evolution of core research themes, identify seminal contributions, and map influential author clusters that have driven innovation in the field. These dimensions are directly related to the study’s core research questions—particularly RQ2 and RQ3—as they reveal the methodological paradigms, disciplinary intersections, and knowledge diffusion patterns that influence how TOD-related challenges are conceptualized and addressed globally. The findings from this section lay the groundwork for understanding thematic clustering and knowledge frontiers in the next analytical stage.

3.2.1. Analysis of Co-Cited Documents

If two documents (or multiple papers) appear in a third document at the same time, then these two documents form a co-citation relationship. Figure 5 shows the network structure of co-cited documents, which shows the important role of these studies in the coordinated development of urban transportation. Through the analysis of co-cited documents, this study has discovered the knowledge structure in the field of sustainable urban transportation from a huge knowledge base, identified key knowledge bases in this field, and understood the development of this discipline. Accordingly, there were 429 co-cited documents in the sample data. In

Table 1. Co-cited documents.

First Author	Year	Freq	Title
Ibraeva A.	2020	53	Transit-oriented development: A review of research achievements and challenges [26]
An D.D.	2019	43	Understanding the impact of built environment on metro ridership using open source in Shanghai [27]
Guo Y.Y.	2020	39	Built environment effects on the integration of dockless bike-sharing and the metro [28]
Gan Z.X.	2020	38	Examining the relationship between built environment and metro ridership at station-to-station level [29]
Ding C.	2019	36	How does the station-area built environment influence Metrorail ridership? Using gradient boosting decision trees to identify non-linear thresholds [30]
Ewing R.	2010	35	Travel and the Built Environment: A Meta-Analysis [10]
Li S.Y.	2020	34	Spatially varying impacts of built environment factors on rail transit ridership at station level: A case study in Guangzhou, China [31]
Li Z.K.	2019	33	Transit oriented development among metro station areas in Shanghai, China: Variations, typology, optimization and implications for land use planning [32]
Chen E.H.	2019	32	Discovering the spatio-temporal impacts of built environment on metro ridership using smart card data [33]

,the top ranked was Ibraeva A. (2020) with 53 co-citations [26], followed by An D.D. (2019) with 43 [27], Guo Y.Y. (2020) with 39 [28], Gan Z.X. (2020) with 38 [29], Ding C. (2019) with 36 [30], Ewing R. (2010) with 35 [10], Li S.Y. (2020) with 34 [31], Li Z.K. (2019) with 33 [32], Chen E.H. (2019) with 32 times [33], and the number of citations for the remaining articles is less than 30 [34,35,36,37,38,39,40,41]. The data shows that most of the literature with a high number of citations is concentrated in 2020, and as the frequency of citations decreases, the time of publication of the literature gradually becomes more diverse, with more of the literature from older years appearing.

3.2.2. Analysis of Co-Citation Author Network

Cooperation between authors in the academic field plays a crucial role in advancing research, as collaborative efforts often lead to more comprehensive studies and innovative solutions. To analyze the productivity and collaborative networks of authors in the domain of urban rail transit and land use, this study utilized CiteSpace software to process and visualize the data. The resulting network of frequently cited authors provides a clear picture of the co-citation relationships that drive research in this field (Figure 6). Authors who are cited more frequently often serve as bridges between different research communities, fostering interdisciplinary dialog and enhancing the exchange of knowledge. According to the network analysis, the leading contributor is Cervero R., with his first collaboration dating back to 2005, and a total of 739 collaborative papers. He is closely followed by Ewing R., with 467 collaborative papers, and Frank L.D., who has co-authored 235 papers. Other notable contributors include Handy S. (166 collaborative papers), and Cao X.Y. (160 collaborative papers). The sixth to tenth positions are occupied by Litmante, Zhao P.J., Hess D.B., Handy S.L., and Debrezion, with 147, 145, 142, 142, and 139 co-authored publications, respectively. These authors have significantly influenced the field through their collaborative work, and their contributions have shaped the current understanding of urban rail transit and land use interactions. This collaborative network underscores the importance of interdisciplinary cooperation in advancing research in this domain.

3.2.3. Analysis of the Co-Cited Journal

The number of documents counted in this study is the total number of citations in the existing literature. It was found that 279 journals published 27,027 articles related to the interdisciplinary field of rail transit and land use. The top ten journals in this research field accounted for 29.74% of total publications, suggesting that academic achievements are concentrated in a few high-impact journals. Notably, Transportation Research Part A: Policy and Practice leads the list with 1029 articles, followed by Journal of Transport Geography with 1005 articles, and Transportation Research Record with 1003 articles. Other prominent journals include Transportation Research Part D: Transport and Environment with 815 articles, Transport Policy with 789 articles, and Urban Studies with 766 articles. The Journal of the American Planning Association and Transportation each contributed 722 and 694 articles, respectively, while Cities and Sustainability published 691 and 524 articles, respectively. Furthermore, journals publishing more than 10 articles in this field account for 75.27% of the total, suggesting that the dataset is highly representative of the journals actively contributing to the discourse on rail transit and land use (Figure 7).

3.3. Keyword Dynamics: Theme Evolution and Emerging Trends

This section focuses on the core knowledge and thematic evolution in this field. Through clustering and analysis of high-frequency keywords, it reveals the stable core (such as built environment, land use, and transportation) and dynamic thematic shifts (such as value capture, gentrification, urban resilience, and machine learning) in the research. This section lays the groundwork for further deconstructing knowledge clusters and constructing a refined thematic structure in Section 4, while also delving into their intrinsic logical connections, interdisciplinary characteristics, and innovation-driven mechanisms.

3.3.1. Keyword Network and Cluster Analysis

A bibliometric analysis of keyword frequencies can reflect research frontiers, as keywords that appear more frequently often indicate widely concerned research priorities. To reflect the research focus over the past 20 years, annual slices were used, and “keywords” were used as the node type. In Figure 8, “land use” is the most important keyword, appearing 666 times, reflecting its central position in academic discussions. The second most frequently occurring keyword is “impact”, which appears 437 times, highlighting the importance of understanding the interaction between urban rail transit and land use. The third most frequently occurring keyword was “built environment,” which appeared 400 times, indicating that scholars generally pay attention to the physical and spatial environment of urban development. This was closely followed by “transportation” and “accessibility,” which appeared 258 times each, indicating that as the built environment around stations improves, stations attract more and more passenger flow, and research on the role of transportation systems in improving urban accessibility has also received increasing attention. In addition, the word “city” appeared 248 times, indicating that research on the environment is mainly focused on large cities. Other keywords appeared less than 200 times, reflecting more specific aspects related to the research topic.

By clustering the frequently occurring keywords, Figure 9 shows seven thematic clusters, and

Table 2. Top terms in clustering.

Cluster	Size	Top Terms (Log-Lihood, Ratio-Level)
#0 built environment	76	built environment (23.55, 0.0001); physical activity (12.47, 0.001); travel behavior (12.47, 0.001); walking (10.41, 0.005); environment design (8.83, 0.005);
#1 value capture	58	non-linear effects (11.36, 0.001); spatial heterogeneity (9.7, 0.005); transit-oriented development (9.68, 0.005); urban design (8.52, 0.005); sustainable mobility (8.52, 0.005);
#2 public transit	53	air pollution (10.3, 0.005); beijing (7.46, 0.01); urban transport (6.86, 0.01); decentralization (6.86, 0.01); urban growth (6.86, 0.01);
#3 china	51	value capture (12.13, 0.001); difference-in-differences (9.09, 0.005); built environment (8.88, 0.005); housing price (8.24, 0.005); transportation (7.94, 0.005);
#4 public transport	21	service (11.38, 0.001); equity (6.08, 0.05); gis (6.08, 0.05); urban vitality (5.68, 0.05); multiscale geographically weighted regression (5.68, 0.05);
#5 travel behavior	5	vulnerability (21.51, 0.0001); resilience (15.36, 0.0001); COVID-19 (15.36, 0.0001); climate change (7.76, 0.01); planning (6.56, 0.05);
#6 transport planning	5	cellular automata (20.87, 0.0001); urban sustainability (14.11, 0.001); system dynamics (11.9, 0.001); transit value capture (7.04, 0.01); difference-in-difference model (7.04, 0.01);

summarizes the thematic clusters and their corresponding top five research focuses. ‘#0 Built environment’ has the longest time span and the largest scale, running through the entire research period. Early terms in this category were relatively broad, covering general concepts such as urban development, urban ecology and land use planning, and included key terms such as ‘physical activity’, ‘travel behaviour’, ‘walking’ and ‘environmental design’. These terms highlight the continued focus of scholars on the relationship between urban form and human mobility. ‘#1 Value Capture’ reflects the land premium behavior brought about by rail transit stations, and includes keywords such as “non-linear effects”, “spatial heterogeneity”, “transit-oriented development” and “urban design” to reflect the increasing research on financing mechanisms with the development of rail transit stations. ‘#2 Public Transport‘ focuses on issues such as “air pollution”, “Beijing”, “urban transport” and “decentralisation”, reflecting scholars’ thinking about the problems of big cities in the process of urban construction, and indicating people’s concerns about the traffic environment and governance level. ‘#3 China’ is characterized by keywords such as “value capture”, “housing prices” and “transport”. China has attracted increasing attention from experts in the fields of urban planning and transport planning due to its rapid urban expansion and continuous construction of transport infrastructure in the past 20 years [42]. As a result, a large number of empirical studies have been carried out on China’s rapid urbanization and transport investment. ‘#4 Public transport’ is characterized by terms such as “service”, “equity” and “urban vitality”, indicating that with urban development and class differentiation, more and more scholars are concerned about how to coordinate transport resources to maintain social equity and the different impacts of public transport facilities on the travel of different social groups. It also further explores the factors that will affect the vitality and healthy growth of cities. ‘#5 Travel behaviour’ is a recent research focus. This cluster includes terms such as ‘vulnerability’, ‘resilience’, ‘COVID-19’ [43] and ‘climate change’, which indicate that under the impact of the epidemic, global crises have an explosive impact on urban travel patterns, and urban resilience has received further attention. ‘#6 Transportation Planning’ indicates that in the face of problems arising from the process of urban development, more and more scholars and government departments have carried out macro and micro adjustments to maintain the sustainable development of cities. This cluster includes terms such as “cellular automata”, “urban sustainability” and “transport value capture”, which reflect advanced modeling methods in transportation policy research.

Figure 9 presents a keyword clustering network where horizontal coloured lines divide the network into distinct thematic clusters, each representing a coherent group of research topics identified through keyword co-occurrence analysis. Within each cluster, curved lines illustrate the co-occurrence relationships between individual keywords. These arcs connect terms that frequently appear together in academic publications, indicating a strong conceptual link. The colour of each curved line corresponds to the cluster in which the connected keywords are located, visually reinforcing the internal consistency of a research theme. When a curved line links keywords across two different clusters, it reflects cross-thematic integration, highlighting the interdisciplinary nature of emerging research. By analysing the horizontal structure and curvature of the connections, it is possible to discern the internal cohesion of each cluster and trace the evolution and convergence of ideas across the broader research landscape. Thus, this dual-layered visualisation reveals both the stability of core themes and the dynamic interplay among them over time. Figure 9 shows that the keyword “United States” appeared early in the research, which is consistent with the results of our statistical analysis of the sources of the literature. During this period, the main focus was on the built environment and TOD, reflecting the initial emphasis on the spatial structure and planning strategies of urban areas. As research deepened, attention turned to urban construction-related issues, as evidenced by keywords such as “density”, “pollution” and “climate change”, which indicate a growing focus on environmental and sustainability issues. Keywords such as “investment,” “housing prices” and “land value capture” emerged during the same period, indicating that the economic and financial impacts of construction and development were more pronounced in areas where rapid rail transit was being built. More recently, the scope of research has expanded to include social and health-related aspects. Keywords such as “older adult,” “green space,” and “COVID-19” indicate an emerging awareness of the role of urban rail transit in promoting public health, accessibility, and response to global crises. In the most recent period, keywords related to emerging technologies such as “big data”, “travel mode choice”, and “machine learning” have emerged, indicating a shift in research focus towards the use of advanced data analysis and AI to optimize transportation planning, predict commuting behavior, and improve decision-making processes.

3.3.2. Research Trend Analysis

In this study, a comprehensive review of the interaction between urban rail transit and land use has been conducted, highlighting key trends and research hotspots over different time periods. Based on the analysis of the provided diagram (Figure 10), the keywords can be categorized into three distinct stages, each representing an evolving focus within the field. The progression of these stages illustrates the changing dynamics of urban rail transit and land use, influenced by technological advancements and contemporary socio-economic developments.

The early emerging keywords include terms that laid the foundation for the relationship between rail transit and urban development. Keywords such as physical activity, bus rapid transit, and TOD dominate this phase. This reflects the fact that the initial phase of research focused on exploring the concept of TOD and the use of high-precision remote sensing data.

In the mid-term, TOD remains an important theme, while at the same time China is receiving increasing attention, reflecting the region’s rapid urbanization and the increase in the mileage of the rail network. The emergence of keywords such as house prices, gentrification and smart card data suggests that much attention is being paid to the socio-economic impacts of rail transport on property values and urban communities. Meanwhile, advanced modeling techniques such as cellular automata methods and feature pricing models have upgraded the original measurements as a tool to quantify the impact of urban rail transit on land use and real estate markets. With the continuous improvement of urbanization in East Asia, keywords such as, China, housing price, gentrification, smart card data, etc., gradually emerged in the middle of the research period. During this period, the level of urbanization in East Asia continued to increase, and the construction of rail transit exacerbated the heterogeneity of urban housing prices, causing more and more middle-class people to gather around rail transit stations. Using advanced modeling methods such as cellular automata, the impact of rail transit on housing prices can be more accurately quantified.

In recent years, major international events and the development of emerging technologies have had a significant impact on academic research priorities. The outbreak of COVID-19 has had a significant impact on urban rail transit operations for some time, prompting academia to focus on research on urban emergency response and resilience. At the same time, updates to machine learning technology have led to the emergence of research hotspots such as subway passenger volume forecasting, and empirical research on non-linear measures has increased significantly.

4. Discussion

Section 3 lays the foundation for Section 4 through keyword clustering analysis, clarifying the research theme and direction. In Section 4, Section 4.1 delves into the bidirectional interactive relationship between rail transit and land use based on the keywords “#0 Built Environment” and “#2 Public Transit” extracted from Section 3. Starting from foundational theory, it analyzes their interaction patterns and corresponding strategies. Section 4.2 combines keywords such as “#1 Value Capture” and “#3 China” to explore land value capture mechanisms and land development strategies. Through practical cases, particularly those in the context of China’s urbanization, it further elucidates the practical application of value capture mechanisms. Section 4.3 corresponds to keywords such as “#4 Public Transport” and “#5 Travel Behavior,” analyzing travel decision-making and commuting mode selection. Section 4.4 corresponds to “#6 Transport Planning” and “RQ3”, reviewing recent technological evolution and key emerging technologies. Overall, a comprehensive research framework with a clear logical structure and seamless continuity is established.

4.1. Bidirectional Interactions Between Urban Rail Transit and Land Use

This section examines the bidirectional relationship between urban rail transit and land use. It focuses on three key aspects: theoretical frameworks, interactive mechanisms and strategic tools for integration. Models such as the Node-Place framework and coupling degree models are introduced to evaluate how transit infrastructure shapes land use patterns, and how land use influences transit performance in turn. Drawing on empirical case studies and emerging technologies such as GIS and AI, the section presents practical approaches to optimizing this coordination. These discussions establish a vital theoretical and empirical basis for subsequent chapters, particularly the analysis of land value capture (Section 4.2) and institutional mechanisms, by clarifying the interaction between physical infrastructure and spatial planning.

4.1.1. Theoretical Foundations of Urban Rail Transit and Land Use

Systematic research on rail transit and land use theory began in the 1960s, with its theoretical evolution following a trajectory from static analysis to dynamic synergy and from a single-perspective approach to multidisciplinary integration. Early studies, represented by Alonso’s rent theory [44] and Hansen’s accessibility model [45], revealed the one-way influence mechanism of transportation infrastructure on land value but failed to adequately consider the proactive role of planning interventions.

Following the widespread adoption of sustainable development concepts in the 1990s, Calthorpe proposed the TOD theory [46], Bertolini developed the node-place model [1], and Smolka [47] introduced the land value capture theory. These frameworks shifted the research focus toward proactive collaborative planning, emphasizing the mutual feedback development of transportation and land through high-density mixed-use development and functional alignment.

In the 21st century, research has further evolved into complex systems analysis, with scholars incorporating complex adaptive systems theory [48,49] and spatial syntax [50], combined with big data technology, to uncover the non-linear synergistic mechanisms underlying multi-stakeholder interactions. In recent years, with the emergence of carbon neutrality goals, low-carbon TOD and equity research have become new focal points, driving theoretical expansion toward sustainability and social inclusion [51]. Currently, how to integrate multidisciplinary perspectives and quantify dynamic synergistic effects remains a key issue requiring breakthroughs in future research.

4.1.2. Dynamic Interactions Between Urban Rail Transit Development and Land Use

There is a complex two-way interaction between rail transit and land use. On the one hand, rail transit stations enhance the accessibility of surrounding areas, promoting land appreciation and functional restructuring, and creating a “corridor effect” whereby commercial and residential land is concentrated around stations, thereby promoting sustainable transportation development [52]. On the other hand, the density and mix of land use influence the scale of passenger flow and operational efficiency at stations, creating a cyclical cumulative effect of “land use—passenger demand—rail service.” For example, Canadian research indicates that the marginal benefits of relying solely on rail transit infrastructure development are limited, but when coordinated with land use policies, it can significantly promote economic growth and achieve TOD objectives [53].

This interactive mechanism is regulated by factors at different scales. At the macro level, urban spatial structure influences the layout efficiency of rail networks, while social factors (such as the concentration of minority groups in employment centers due to transportation restrictions in the United States) also affect passenger flow distribution [54]. At the meso level, the intensity of planning controls in station areas (e.g., zoning reforms and TOD incentives in New York and Toronto) affects the degree of development coordination, while the alignment of transportation corridor planning with regional density directly impacts rail transit utilization rates [55]. At the micro level, built environment design (e.g., “last-mile” connectivity experiences) and mixed-use land use integration [56] influence residents’ mode choice, thereby affecting passenger flow generation.

The synergistic development of urban rail transit and land use exhibits significant spatial variations and dynamic changes. For example, cities in Europe and America (such as Barcelona [57] and Stockholm [58]) and Asian cities (such as Shanghai [59] and Tokyo [60]) exhibit different passenger flow-land use coupling patterns due to differing land policies. Additionally, technological advancements (such as spatial measurement methods and machine learning) can optimize transportation network design [61], but issues such as transfer system optimization [62] and social equity (e.g., accessibility for low-income groups) still require consideration. Therefore, future research should establish a more refined analytical framework, combining quantitative models (e.g., site selection simulation) [63], policy evaluation, and social equity analysis to comprehensively reveal their complex interplay.

4.1.3. Strategies for Enhancing Sustainable Rail Transit-Land Use Synergy

Recent developments in data-driven techniques offer fresh opportunities to improve the coordination between rail transit and land use. Geographic Information System (GIS)-based accessibility models refine spatial analysis by charting transit service areas and walkability measures, providing valuable, data-informed insights for policies promoting transit-oriented development [64].

In addition, analyzing mobile phone data helps to monitor passenger flows and travel demand in real time, which helps to strategically locate stations and design intermodal connections [65]. In addition, by integrating AI and machine learning into planning and design, non-linear passenger flow and land use prediction models are built to help urban planners manage potential development challenges [17].

Synergistic development between rail and urban land use can improve station accessibility, reduce congestion and support efficient land development. Key ways in which the two interact include mechanisms such as land value capture, land use regulations and governance reforms. Surplus land values from public transport investments are reinvested by refinancing rail infrastructure through tax improvements, tax incremental financing (TIF) and public–private partnerships (PPPs) [66]. The Urban Development Act encourages high-density and mixed-use development around stations by concentrating foot-traffic-concentrated public buildings as well as commercial areas around rail stations to increase the level of passenger supply on the one hand and reduce the demand for traveling by other modes on the other hand, slowing down urban sprawl [26]. Effective planning requires close coordination and integration between different sectors of government and society [67]. Linkages between stations and neighboring areas can be strengthened through the establishment of a comprehensive governance framework and improved financing models for land development. Future research should focus on AI-driven urban big model planning, integration of multiple types of intermodal transport systems, and summaries of empirical studies in other parts of the world.

4.2. Mechanisms of Land Financing and Value Capture

This section systematically explores the application logic of land value capture mechanisms in urban rail transit financing. By analyzing the relationship between urban rail transit investment and land value appreciation, the study reveals how governments can recover value through taxes, transfer of development rights (TDR) and joint development strategies. The final section focuses on the conflicting interests of governments, developers and the public, emphasizing the need to establish a fair value-sharing framework.

4.2.1. Value Capture

Value capture refers to the process by which governments recover land value gains generated by transport investments through mechanisms such as taxation and the TDR. The revenue collected is then reinvested to subsidize public transportation systems [68,69]. The core principle underlying this approach is that rail transit enhances land value by improving accessibility, creating a transmission chain of “accessibility → land value → capitalization” [70].

The expansion of urban rail transit networks has a multidimensional impact on urban land use patterns and spatial structures. Research indicates that rail infrastructure directly influences land development intensity and housing prices [71], as well as driving urban spatial restructuring by attracting the aggregation of commercial activity. For instance, the construction of the Sapporo subway system in Japan resulted in a substantial increase in the city’s built-up area [72]. Within an 800-meter radius of newly constructed subway stations in the outskirts of Beijing, the catering and services sector exhibited a trend towards concentration, while also stimulating the demand for integrated development of underground commercial spaces, parking facilities, and public areas. Cities such as Tokyo, Hong Kong and Shanghai have improved both the travel experience of rail passengers and the efficiency of land utilization by constructing multifunctional underground systems. However, empirical research in Los Angeles suggests that incompatible zoning policies may restrict the development potential of land around subway stations, thereby suppressing the economic growth triggered by rail transit [73].

The conclusion that “rail transit significantly influences surrounding land values” has been widely validated by the academic community. Following the expansion of the metro system in Sunderland, UK, property prices around stations showed significant appreciation [68]; research in Wuhan, China, indicates that mixed-use areas exhibit the most pronounced property value appreciation effects [74]. Research based on complex network theory further reveals that the number of transportation nodes, network clustering coefficients, and land prices are positively correlated, while path length is negatively correlated with land prices [75]. These findings provide theoretical support for land value capture policies and reinforce the practical logic of “using rail transit investment to drive land appreciation gains that are reinvested in infrastructure development.”

4.2.2. Mechanism Anatomy of Land Financing

Land financing mechanisms are a core tool for urban spatial design; governments around the world commonly adopt land value capture strategies to recover the spillover benefits of transport infrastructure and convert them into reinvestment capital for public projects. Among these, Transfer of Development Rights (TDRs), Tax Increment Financing (TIF), and Joint Development are three mechanisms that are widely used in countries around the world.

The TDR mechanism establishes a development rights trading market, allowing land rights holders in areas subject to development restrictions to transfer their unused floor area ratio to high-density development areas. This system not only protects historical heritage and ecological space, but also guides the orderly expansion of urban space. Taking the American zoning control system and the Italian historic urban renewal plan as examples, the TDR mechanism has successfully achieved the coordinated governance of land resource protection and urban density improvement [76]. On this basis, empirical studies have shown that the regulated operation of the TDR market can not only avoid the pressure of new tax burdens, but also effectively leverage private capital to participate in the urban development process, forming a financing model of co-governance by multiple entities [77].

The policy effectiveness of the incremental tax financing mechanism is mainly reflected in the fields of urban renewal and infrastructure financing. TIF enables municipalities to reinvest future tax revenues generated from increased property values into infrastructure improvements [78]. A case study from Fort Worth illustrates how TIF revenues were successfully employed to revitalize historic districts, transforming them into residential and commercial hubs [79].

Joint development is an essential strategy in TOD financing, fostering collaboration between governments and private developers to co-finance infrastructure projects. Under this model, governments provide land-use incentives (such as density bonuses or direct development rights) in exchange for private investment in public transit or urban renewal projects [80]. Case studies from Hong Kong’s MTR and Tokyo’s railway-led urban development demonstrate that aligning real estate projects with transport infrastructure investments enhances financial sustainability and urban connectivity [81]. However, researchers emphasize that clear governance frameworks and equitable revenue-sharing models are necessary to mitigate risks and ensure public benefits [77].

Future research on land mechanisms should focus on the dual-wheel drive of institutional innovation and technological empowerment. Mechanism adaptability should be deepened through cross-institutional environmental comparisons and multi-dimensional integrated design. AI dynamic models should be combined to achieve precise governance. At the same time, attention should be paid to social spatial justice, and a resilient urban development framework that balances the efficiency of land capitalization with social equity should be constructed. This will provide a systematic policy toolbox for addressing climate change and urban renewal.

4.2.3. Key Conflicts: Government–Market–Public Game in Value Distribution

The LVC mechanism is a crucial tool for financing urban development, yet it often generates conflicts among three major stakeholders: the government, the private sector, and the public. These conflicts are particularly pronounced in China’s state-owned land system, where the government holds monopoly control over land supply and uses land-based financing to fund infrastructure and economic development [82]. While LVC strategies such as public land leasing, taxation, and joint development projects can generate substantial revenue, they also exacerbate social inequalities, leading to disputes over income distribution, property affordability, and urban redevelopment priorities.

From a government perspective, land sales and leasing are major revenue sources for local authorities, which depend on these funds for infrastructure expansion and debt servicing [83]. However, excessive reliance on land-based financing has led to soaring property prices, placing home ownership beyond the reach of middle- and low-income groups [84]. The government’s dual role as both a regulator and a land seller creates conflicts of interest, as policies intended to curb speculation often contradict the need for sustained revenue generation.

The private sector, particularly real estate developers, advocates for reduced government intervention to facilitate market-driven development. However, developers tend to prioritize high-end housing projects, contributing to land supply distortions and reducing the availability of affordable housing [85]. In some cases, PPPs have attempted to balance market efficiency and public benefit, as seen in Hong Kong’s MTR rail-plus-property model and Tokyo’s railway-led urban development. However, without equitable benefit-sharing mechanisms, private developers often capture a disproportionate share of land appreciation gains, leading to public discontent [82].

The public, particularly lower-income residents, faces increasing housing affordability challenges due to rising land values. Public opposition to land expropriation, speculative investments, and urban renewal projects has become more vocal, especially in cities like Guangzhou and Shenzhen, where disputes over compensation for displaced residents have led to protests [86]. The need for inclusive development policies, such as inclusionary housing and social rental housing programs, has become more pressing.

To address the conflicts arising from LVC, it is essential to establish a comprehensive and transparent land value redistribution framework. Policymakers should focus on three key areas. First, strengthening LVC regulations by introducing progressive land taxation policies and equitable redistribution mechanisms can ensure that land value gains benefit governments, developers, and the public alike. Second, balancing market efficiency and social equity through inclusive housing policies, such as affordable housing quotas and community benefit agreements, can mitigate market-driven inequalities and enhance public well-being. Third, increasing public participation in urban governance by establishing community engagement platforms will allow local voices to influence planning decisions, fostering a more democratic and transparent land development process.

Future research should focus on resolving conflicts between government, market forces, and public interests through interdisciplinary approaches. Key areas of exploration include redefining the government’s role by separating land ownership and regulatory functions, innovating market mechanisms through dynamic value-added income distribution models, and integrating digital public participation tools such as blockchain contracts and digital twin negotiation systems. Additionally, researchers should work on quantifying spatial justice using three-dimensional land price modeling and AI-driven site selection algorithms for affordable housing. Comparative studies across China’s city policy database will contribute to a political economy framework tailored to China’s unique land governance system. Finally, urban experimentation through policy laboratories will provide empirical insights into balancing efficiency and fairness in urbanization strategies.

4.3. The Impact of the Built Environment on Travel Behavior

This section focuses on the keywords ‘0# built environment’ and ‘5# travel behaviour’ in keyword clustering and establishes a three-layer analytical framework. Firstly, beginning with spatial attributes such as floor area ratio and land use patterns, we clarify how these attributes, alongside individual socioeconomic characteristics, influence commuting patterns. This verifies the correlation between the two themes. Secondly, within the TOD framework and incorporating behavioral economics theory, we analyze the impact of economic costs, spatial form and external shocks on travel decisions. Thirdly, we strengthen the logical connections through technical validation by integrating data from both sources using technologies such as big data and AI modeling. This lays the groundwork for further research on technology-enabled TOD models and sustainable travel strategies in future smart city development.

4.3.1. The Role of the Built Environment in Mobility Choices

Numerous empirical studies have shown that urban plot ratio, land use patterns, transport interchange patterns, and the distribution and quality of transport facilities play a crucial role in the use of transport modes [87]. Studies have shown that the socio-economic characteristics of individuals and households [88] and the spatial attributes of the built environment [89] together shape commuting patterns and overall travel demand. In particular, good neighborhood planning, high-density urban development and a well-developed street network can significantly promote the use of walking, cycling and public transport, thereby influencing residents’ travel mode choices. Well-designed urban areas with integrated transport systems encourage active travel modes, while poorly planned environments discourage non-motorized travel and increase car dependence [90]. These findings emphasize the importance of linking urban design and transport policy for sustainable travel.

The extent of urban sprawl and the pedestrian environment are also key factors influencing travel. A study in Las Vegas found that unconstrained urban sprawl reduces the ease of walking [91]. Research in Wellington, New Zealand, found that land use and transport policies influence active commuting behavior, with GIS and WILUTE modeling demonstrating that increasing bus frequency and reducing fares significantly increased walking and cycling rates [92]. In Beijing, the large number of bike-sharing bikes around underground stations has improved the ‘last mile’ problem, expanding the underground traveling distance and the number of public transport riders [93]. A study using graphical Fourier transform theory found that bicycle-friendly transport corridors in Seoul reduced rail congestion, illustrating the complementary relationship between bicycles and public transport [94].

4.3.2. Travel Decisions and Commuter Mode Choices

Commuting mode choice is an important aspect of urban transport planning and is influenced by a combination of economic, spatial and behavioral factors. Within the framework of TOD, commuting decisions are influenced by multiple factors such as economic conditions, transport policies, and infrastructure investments [95].

Behavioral economics has highlighted the role of travel costs and economic incentives in commuting mode choice, especially among low-income populations [95]. The TPB research has shown that travel costs, convenience and personal attitudes directly affect travel mode preferences, with people relying less on public transport and using active transport or cars when fares increase. A case study in Wellington, New Zealand, showed that lower fares and increased frequency of service on public tools significantly increased ridership and active commuting behavior [92]. Urban form also plays an important role in influencing commuters’ preferences and multi-modal transport switching. Studies have shown that the centrality of the metro network directly affects commuting patterns, and that transit nodes that integrate multi-modal transport minimize transfer costs and improve overall commuting efficiency [4]. A study of integrated transport nodes in Europe has shown that public transport utilization is higher in cities that have effectively combined selected multi-modal transport options such as rail, buses and microtransit [90]. In Germany, where a public transport fare-free scheme was carried out for a period of time, it was found that such policies could temporarily increase ridership but were not very effective in shifting transport choices in the long term and required more urban planning design interventions [96]. In the current phase of natural disasters and extreme events, the resilience of transport infrastructure is increasingly recognized as a key influence on commuting mode choice. When social or natural emergencies occur, disruptions in the public transport network can lead to a significant shift towards the use of private vehicles, increasing congestion and environmental burdens [97]. These findings highlight the importance of resilient public transport planning for sustainable urban transport systems. Technological advances such as mobile phone signaling data, AI-supported traffic modeling and multimodal integration strategies are gradually transforming transport planning [68]. Combining big data analytics and AI to predict commuting trends optimizes public transport pricing and infrastructure investment.

4.3.3. Technological Innovations in Mobility Planning

Advances in big data analytics, AI and Machine Learning can optimize the way cities plan transport and operate their public transport systems. New technologies such as real-time traffic monitoring, predictive analytics and demand response have greatly improved the efficiency and accessibility of traffic coordination. Artificial intelligence-driven multi-modal transport macromodels have shown that these models are able to analyze complex travel behavior patterns and provide more adaptive and humane transport solutions [98]. A study of the Beijing Metro found that the application of mobile phone data and machine learning algorithms to analyze origin–destination (OD) passenger flows provided deeper insights into commuter mobility trends and improved the accuracy of transit demand forecasts [99].

TOD strategies have been improved through technological advances that have increased the level of coordination between land use and transit integration models. Research on GIS has shown that remote sensing technology is an artificial intelligence tool that can now be used to manually model the impact of accessibility and optimize the location of transit hubs through satellite imagery [100]. Research in digital twin technology further supports the implementation of TOD planning by creating real-time urban simulations that help decision makers predict the long-term effects of public transport investments [101]. These innovations contribute to evidence-based decision-making, reduce congestion, improve multimodal transport and encourage sustainable urban development.

The deployment of autonomous vehicles (AVs) and shared mobility is another emerging trend. A study on AV integration in Singapore found that strategically deploying AVs within first-mile and last-mile networks enhances public transit ridership and accessibility [102]. However, research suggests that unregulated AV expansion could undermine public transport systems, potentially increasing congestion and emissions unless carefully managed through policy interventions and smart urban planning [103].

Advancements in AI, ML, and big data analytics are reshaping mobility planning and TOD implementation, offering new opportunities for sustainable, data-driven urban transport solutions. Future research should explore AI-powered transit management, digital twin urban modeling, and climate-adaptive transport policies to ensure efficient and resilient urban mobility systems.

4.4. Key Techniques: Categories, Applications, and Limitations

4.4.1. History of Technological Development

In the early 2000s, urban rail transit and land use integration began with the rise of 3S technologies: GIS, GPS and RS, GIS provided the spatial foundation for land-use zoning, buffer analysis, and route planning [104,105]. RS allowed for large-scale monitoring of urban sprawl and detection of built environment changes through satellite imagery [106,107]. GPS data supported travel pattern studies and informed accessibility models by pinpointing real-time transit usage [108]. 3S technology has initially established a technical framework for research on the integration of urban rail transit and land use, addressing fundamental issues such as spatial analysis and data acquisition. However, it is difficult to explore the complex interactive relationship between the two. As research deepens and faces more complex interactions between land use and transportation, the data foundation provided by 3S technology alone is insufficient to meet the demand, leading to the emergence of spatial modeling tools.

As urban datasets expanded, spatial modeling tools evolved. LUTI models became increasingly influential, simulating how land development intensity affected ridership, congestion, and housing prices [14]. Researchers introduced Node–Place models to balance land intensity (“place”) with transport connectivity (“node”), which informed optimal site selection for transit hubs [1,109,110]. With the digitization of transport systems, researchers began leveraging smart card data, mobile phone GPS, and POI data to assess urban mobility and land use on a granular level [111]. Transit authorities used this information to derive OD matrices and visualize passenger flows in real time [112]. Additionally, cellular automata, system dynamics models, and real-time geospatial platforms were deployed to simulate urban evolution, providing interactive decision tools for planners [113]. These innovations allowed for dynamic modeling of traffic, land value fluctuations, and housing demand near stations. Spatial modeling tools effectively simulate the mutual influence between land use and transportation by constructing various models, thereby optimizing the layout of transportation facilities.

The most recent phase has seen a surge in AI-powered urban analytics, with machine learning applied to forecast transit ridership, predict land values, and detect gentrification risks around station areas [114]. Deep learning models have further enhanced the accuracy of land-use classification and mode choice prediction [115]. At the infrastructure level, digital twins replicate entire cities—including rail lines, buildings, and demographic flows—within 3D, real-time models that support predictive simulation and crisis management [116]. These twins integrate data from IoT devices, BIM models, and transit APIs, enabling real-time planning for congestion, disaster response, and land policy optimization [117]. In parallel, smart city platforms are operationalizing these tools into municipal planning processes, making transit-land coordination more intelligent, equitable, and adaptive [118]. Artificial intelligence and digital twin technologies have enabled intelligent prediction and management of urban transportation and land use, greatly improving the scientific nature and foresight of decision-making, and promoting the integration of urban rail transit and land use to a higher level. Future research should focus on integrating AI ethics, open-source standards, and community engagement platforms to foster more inclusive and resilient cities.

4.4.2. Key New Technologies

In recent years, urban rail transit planning and land use coordination have increasingly relied on computational intelligence, driven by the rapid evolution of artificial intelligence, data science, and simulation modeling. These technologies not only provide analytical rigor but also enable predictive modeling, scenario simulation, and real-time decision support (

Table 3. Key techniques: categories, applications, and limitations.

Technology	Key Methods	Applications	Strengths	Limitations
Machine Learning	Random Forest, XGBoost, Support Vector Machines, K-means Clustering, Latent Dirichlet Allocation	Land use classification, identifying spatial patterns	High accuracy, scalable automation, effective for classification and clustering tasks	Declines in heterogeneous environments, poor model interpretability
Deep Learning	Convolutional Neural Networks (CNNs) Generative Adversarial Networks, Graph Neural Networks (GNNs)	Zoning analysis, simulating urban form evolution, modeling topological relations	Handles complex and high-resolution spatial data, high-precision spatial analysis	Requires large labeled datasets, high computational resource demands
Evolutionary and Swarm Intelligence	Genetic Algorithms (GA), Particle Swarm Optimization (PSO), MOEA/D, NSGA-II	Land use configuration optimization, transit-oriented development layout planning	Good at solving multi-objective optimization, effective in exploring planning scenarios	Cannot dynamically respond to real-time changes in land markets, travel demand, and zoning regulations
Simulation Models	Cellular Automata Markov Chain Models	Urban expansion projection, land use–transport interaction simulation	Improved accuracy when considering spatial interaction	Lack of responsiveness to real-time fluctuations

).

Machine learning has emerged as a fundamental tool in urban predictive modeling, particularly for land use classification, mobility pattern recognition, and demand forecasting [119]. Supervised learning models, including Random Forest, XGBoost, and Support Vector Machines, demonstrate high accuracy in predicting land use categories and estimating transit hub passenger flows [120]. Unsupervised techniques such as K-means clustering and Latent Dirichlet Allocation effectively identify latent spatial patterns in Points of Interest datasets and geotagged social media data [121]. While ML enhances automation and scalability in urban analytics, its performance may decline in heterogeneous environments, and model interpretability remains a persistent challenge.

Deep learning excels in processing complex spatial data and high-resolution imagery. Convolutional Neural Networks extract fine-grained spatial features from satellite and street-view imagery, supporting zoning analysis and transit-oriented development mapping [122]. Generative Adversarial Networks, exemplified by LUCGAN+, simulate urban form evolution under diverse planning scenarios [123]. Graph Neural Networks model topological relationships between transit nodes, pedestrian networks, and land parcels [124]. Despite enabling high-precision spatial analysis, DL methods demand extensive labeled datasets and significant computational resources.

Evolutionary and swarm intelligence algorithms address complex trade-offs in land allocation. Genetic Algorithms and Particle Swarm Optimization optimize land use configurations under accessibility, ecological, and compactness constraints [125]. Advanced techniques like MOEA/D and NSGA-II enable high-resolution exploration of transit-oriented development layouts [126]. Hybrid approaches can effectively simulate the long-term interactions between land use and transport. Cellular automata and Markov chain models project urban expansion under different planning scenarios [127], while simulation results generated by the cellular automata model considering spatial interaction are significantly improved [128]. However, a key limitation is their inability to respond dynamically to real-time fluctuations in land markets, travel demand and zoning regulations.

5. Conclusions and Future Prospects

5.1. Research Progress and Existing Challenges

In the past two decades, with the acceleration of global urbanization and technological innovation in analysis, research on the collaborative mechanism between urban rail transit (URT) and land use has formed a systematic theoretical framework. This systematic review shows that (1) the research network shows significant geographical agglomeration characteristics, with China and the United States as the core nodes of knowledge production, forming a global academic cooperation network; (2) research mainly focuses on three major areas: the spatial feedback mechanism between the built environment and land use, the capture of rail transit premiums and innovation in financing models, multi-modal transportation network and travel decision behavior research; (3) the methodology has evolved through three stages: a transition from the neoclassical ground rent theory model to the spatial econometric model. In recent years, through deep integration with machine learning algorithms and smart city simulation technology, a comprehensive analysis method for multi-scale and multi-source data has been formed.

Current research still has three limitations: First, existing results are mostly concentrated on developed urban agglomerations in the northern hemisphere, with insufficient attention paid to urbanized areas in the southern hemisphere. Second, the dynamic complexity caused by the spatial expansion of megacities urgently requires the development of an integrated model that takes into account both macrostructural regulation and microbehavioral simulation. Third, there are institutional barriers to the data acquisition mechanism, and the phenomenon of data silos across departments and fields restricts the depth of research. A sustainable cross-institutional data collaboration platform needs to be established to ensure that data infrastructure keeps pace with technological development.

5.2. Future Research Directions and Implementation Paths

(1)

Technology integration: Researchers should develop a decision support platform that integrates artificial intelligence and urban systems engineering. By integrating multi-source and heterogeneous data streams such as Internet of Things infrastructure data and social media behavior trajectories, an intelligent model with a dynamic feedback mechanism can be constructed. The focus is on breaking through the limitations of traditional machine learning methods in representing spatial heterogeneity and achieving a transition from static correlation analysis to dynamic analysis.

(2)

Data governance: Establish a data collection framework for two-way calibration of “macro policies-micro behaviors”. At the macro level, strengthen the standardized processing of new types of geospatial data such as satellite remote sensing and mobile phone signaling. At the micro level, the creation of a dynamic database, including behaviors such as commuting preferences and land transaction decisions, enables the secure integration of cross-domain data while ensuring that data security and privacy are effectively protected.

(3)

Institutional innovation: Promote the formation of a collaborative research mechanism across disciplines and sectors, and establish a multi-actor participation mechanism covering government, businesses, and communities to find solutions to the challenges of spatial rights and interest allocation.

By addressing these issues, research in the field of rail transit and land use can transition from descriptive to normative and actionable, enabling cities to use urban transport systems as a catalyst for an equitable, sustainable, and resilient urban future.