Liquefied natural gas (LNG) shipping is a sensitive barometer of geopolitical shifts: changes in shipping routes between ports reveal how trade alliances are being restructured. We propose a higher-order temporal network framework to track cooperation between global LNG ports and quantify how these relationships respond to major events. Using global vessel trajectory data from 2022 to 2023, we infer port-level cooperation trends, track the formation/dissolution of port clusters and contract renewals. The results show that 88.2% of countries renewed existing partnerships, and stronger cooperation intentions contribute to stabilizing port cluster structures. We propose the RPI index, which shows high correlation with the Energy Security Index in developed countries. The RPI quantifies the extent to which countries are affected by the Russia-Ukraine conflict, with countries having military/trade cooperation with Russia or developing countries adjacent to the US exhibiting extremely high RPI values. It can also be used to illustrate the division of labour among ports within the same country. Furthermore, the results indicate that exporting countries are seeking new markets; the US and Russia increased their connections with new partners by 8.2% and 2.7%, respectively, while transshipment patterns in Europe also changed, particularly in Spain and France. Overall, the proposed framework provides an event-aware perspective for detecting restructuring in the LNG supply chain and offers quantitative support for policy making in the global LNG shipping network.

Carsharing provides a convenient travel option and has attracted increasing attention in Chinese market in recent years. This study conducted an analysis of carsharing travel patterns and purposes across multiple Chinese cities. We collected over 4.8 million trips from 58 cities, employed Bayes’ rules to infer users’ trip purposes and analyzed their travel patterns within a comparative analytic framework. The analysis explores variations in carsharing usage and trip purposes across cities and examines how urban characteristics influence various trip purposes. The results reveal a strong preference for leisure-oriented travels using carsharing services. There are more work-related trips during daytime and leisure activities in the evenings on weekdays and predominantly leisure activities on weekends. Additionally, there is considerable variability in the distribution of trip purposes across cities. The megacity demonstrates a more balanced usage pattern, whereas the Super-Large City shows a higher demand for commuting trips. In Large City I and Large City II, carsharing is primarily used for leisure and recreational trips. Some carsharing system characteristics, including the number of carsharing parking spaces and shared vehicles, are associated positively with carsharing utilization for different trip purposes, while other factors, like population and road densities, show varying impacts. This study offers a systematic exploration of carsharing services in China and contributes to improving the efficiency of carsharing services in diverse urban contexts.

Ensuring equal access to urban parks is vital for sustainable city development. However, prevailing assessment of park accessibility, which mainly rely on spatial proximity, offer a limited perspective by overlooking destination appeal and journey quality. Utilizing social media and street view images, this study proposes a novel multimodal framework that advanced the evaluation of park accessibility from spatial proximity to experiential quality. The framework first integrated online public feedback and offline facility quality using the PCA-AHPTOPSIS method to evaluate park attractiveness. Then a greenery-perceived distance was proposed by integrating street-level visual greenness and travel distance across three modes to model the perceived impedance of journeys. The effectiveness of the proposed framework is tested in the main area of Wuhan, China. The results reveal that the proposed attraction coefficient correlates more strongly with actual park usage (R2 = 0.68) than area-based metrics (R2 = 0.41). Incorporating GVI significantly improves the explanatory power for visitation frequency, particularly for walking (R2 increase from 0.45 to 0.50) and cycling (from 0.44 to 0.48). GVI's influence varies across travel modes: within a 30-min threshold, walking visitation shows a strong positive correlation with GVI while being weakly affected by distance, whereas distance remains the dominant deterrent for cycling and driving. A pronounced spatial disparity is identified, with 45.3 % of walking zones lacking park access within 30 min, highlighting a mismatch between accessibility and population density. This framework provides practical tools for urban planning and public health, contributing meaningfully to the long-term sustainable development of cities.

Understanding the fluctuations of cross-city travel demand is crucial for the efficient operation of intercity transport systems. Public events, such as concerts and fireworks displays, can cause irregular surges in cross-city travel demand, leading to potential overcrowding, travel delays, and public safety concerns. To better anticipate and accommodate such demand surges, it is essential to estimate cross-city visitor flows with awareness of public events. While prior studies primarily focused on the impacts of a single mega event or disruptions around an individual venue, this study extends the scope by proposing a generalizable framework to analyze visitor flows under multiple concurrent events. We propose to leverage large language models (LLMs) to extract event features from multi-source online information and massive user-generated content on social media platforms. Specifically, social media popularity metrics are designed to capture the effects of online promotion and word-of-mouth in attracting visitors, with their effectiveness verified through comparative analysis. An event-aware machine learning model is then adopted to uncover the specific impacts of different event features and ultimately predict visitor flows for upcoming events. Using real-world events, social media, and visitor arrival data from Hong Kong, the framework is applied to predict daily flows of Chinese Mainland visitors, achieving a testing R-squared of over 85%. We further investigate the heterogeneous event impacts on visitor numbers across different event types and major travel modes. Both promotional popularity and word-of-mouth popularity are found to be associated with increased visitor flows, but the specific effects vary by the event type. This association is more pronounced among visitors arriving by metro and high-speed rail, while it has less effect on air travelers. Through a case study of a specific venue, we demonstrate how event-aware visitor flow prediction facilitates coordinated inter-agency measures and guides the development of specialized transport policies. Such analysis enables the implementation of targeted policies, including responsive operations of border control, dedicated shuttle services to event venues, and comprehensive on-site traffic management strategies.

Bike sharing systems support sustainable urban development, with accurate demand prediction being essential for efficient operations. Previous studies have primarily modeled spatial dependency of bike sharing demand in Euclidean space or among bike stations, but often overlooked topological dependency of demand shaped by urban transportation networks. Metro and cycling networks could influence bike sharing usage through their functional connections with bike sharing systems. To address this gap, this study proposes GeoTopo-Net, a novel deep learning framework to improve short-term demand forecast for urban bike sharing systems. Different from existing solutions, GeoTopo-Net jointly models dependencies of travel demand in both continuous and network spaces. The model utilizes convolutional neural networks (CNNs) to capture spatial dependency between urban areas and their surroundings, while integrating graph convolutional networks (GCNs) to model the topological dependency introduced by urban transportation networks. Our evaluation across five global cities shows that GeoTopo-Net significantly reduces prediction errors, by up to 8.9% in RMSE, 6.8% in MAE, and 5.9% in MAPE. Incorporating dependencies from metro networks produces notable improvements in high-demand areas and those near the metro stations. These findings highlight the importance of incorporating urban transportation network structures in bike sharing demand forecast. The GeoTopo-Net architecture can also be adapted to improve short-term forecast for different types of travel demand (e.g., ride-hailing; electric vehicle charging demand) that involve complex interdependencies in continuous and network spaces.

Proximity-based land use planning has traditionally been viewed as a key strategy to manage travel demand and promote sustainable mobility. Yet, the rise of online activities, particularly the spread of e-shopping, may alter this rationale by reducing the need for physical proximity to consumption opportunities and potentially reshaping residents’ interactions with neighborhood facilities. In this study, we draw on a one-month mobile signaling dataset from Shanghai, China, to provide large-scale behavioral evidence on this issue. Leveraging a Spatial Error Model (SEM), we examine whether e-shopping attenuates the influence of neighborhood land use features on shopping travel behaviors. Our findings show that residents in neighborhoods with higher levels of e-shopping engagement tend to make fewer offline shopping trips and travel shorter distances. More importantly, e-shopping moderates — and in many cases attenuates — the effects of key proximity-based features, including proximity to commercial centers, geographical location, road density, and transit accessibility. We also find significant heterogeneity across neighborhood types: communities with a higher share of elderly residents show greater sensitivity to these moderating effects, particularly in terms of shopping trip frequency. Overall, the results suggest that e-shopping may reshape the foundations of proximity-based planning. Our study further demonstrates the value of mobile signaling data in capturing these emerging dynamics, offering new insights for modeling travel demand and informing land use policies in the digital era.

Trajectory Generation (TG) enables realistic simulation of individual movements for applications such as urban management, transportation planning, epidemic control, and privacy-preserving mobility analysis. However, existing TG methods, particularly unconditional diffusion models, struggle with spatiotemporal fidelity as they often overlook some travel patterns that are critical in an individual's mobility behavior, such as recurrent location visits, movement scope, and temporal regularities. In this work, we propose the Autoregressive Diffusion Model for Travel-Pattern Aware Trajectory Generation (Traveller), a novel approach that integrates autoregressive travel-pattern modeling (AR-TempPlan) with diffusion-based trajectory generation (TravCond-Diff) to produce realistic and context-aware movement patterns. By leveraging the spatial anchor and temporal modes of visiting different locations, we derive an individual's particular travel pattern as spatiotemporal constraints for guided trajectory generation. Building on this, AR-TempPlan generates a mask location sequence as the temporal modes, planning location transitions over time, while TravCond-Diff leverages this planning signal and home location, the spatial anchor, to guide spatial generation through a discrete diffusion process. Experiments on real-world datasets demonstrate that Traveller with the dual guidance mechanism enables the production of high-fidelity and individual trajectories that effectively capture complex human mobility behaviors while preserving privacy. The code and data are available at https://github.com/YuxiaoLuo0013/Traveller.

Shared autonomous vehicles (SAVs) are expected to enhance urban transportation efficiency through innovative mobility resource management. By developing a comprehensive agent-based simulation framework, this study investigates several key factors influencing fleet size and parking demand for the adoption of SAVs in future urban mobility systems. The framework evaluates how both operational factors (e.g., reservation time and maximum waiting time) and demand-side characteristics (e.g., demand rate and the balance between trip origins and destinations) jointly affect the performance of the SAV system. It uses a two-stage simulation process that includes capacity estimation and performance evaluation. In the initial warm-up stage, the simulation estimates the fleet size and parking spaces needed to serve specific travel demand. These initial estimates are then used in the second stage to run further simulations and assess additional performance indicators, including final required parking spaces, empty meters traveled, and trip rejection rate. To obtain a holistic understanding of the studied factors, we construct various simulation scenarios based on historical taxi data in central areas of Chengdu, Shanghai (China), and Manhattan of New York City (USA), and build structural regression models on the simulation outcomes. The results reveal a general mechanism by which operational characteristics and demand patterns influence SAV fleet and parking sizes. We find that a 1 % increase in overall travel demand results in about a 1 % increase in the number of SAVs needed and required parking spaces. Meanwhile, a 1 % improvement in the balance of origin-destination (OD) trips, which reduces spatial mismatches between vehicle supply and trip requests, can help offset the need for additional vehicles and parking spaces. These findings offer critical policy implications, emphasizing the need for integrating SAV deployment with land-use strategies, balancing fleet investment, environmental costs, and service quality (e.g., lower waiting time) in SAV planning and operations.

Although “segregation” is widely used and often treated as a clear condition, what it actually means for a place to be “highly segregated” remains ambiguous. Studies invoke the term across diverse social and spatial contexts, from income to race, or from residence to facility access, without a shared definition of what segregation entails. This ambiguity obscures comparisons across cities and dimensions, risks misdiagnosing the sources of inequality, and undermines targeted interventions. To address this conceptual and empirical gap, we propose a socio-spatial dimensional framework that defines segregation as the intersection between social dimensions (who differs) and spatial axes (how that difference is structured and experienced). The social dimensions correspond to the Social Vulnerability Index (SVI), including Socioeconomic Status (SES), Household Composition & Disability (HCD), Minority Status & Language (MSL), and Housing Type (HT). The spatial axes capture metropolitan unevenness and local clustering. Using mobility data from 45 million mobile devices, we construct this mobility-based framework, Social Segregation Index (SSI), for the 30 largest U.S. metropolitan areas. Results show that “high segregation” has no single meaning: rankings of metropolitan areas change dramatically by social dimensions, with severity and clustering strongest along SES and MSL but weaker for HCD and HT. Extreme multi-axis segregation is rare yet concentrated in a few urban cores. By clarifying who differs and how in space, our framework transforms the vague notion of segregation into a measurable, comparable, and policy-relevant socio-spatial condition.

The rapid growth of tourist-generated content demands scalable and reliable quality assessment methods. This study introduces an LLM-driven framework that combines parameter-efficient finetuning and prompt engineering to evaluate content quality accurately and interpretably. Applied to 484,930 reviews from MaFengWo, TripAdvisor, and Ctrip, the approach achieves superior performance (RMSE=0.3040, NDCG@100=0.500, BERTScore=77.95%) with 9 x higher efficiency. Spatial-temporal-semantic analyses reveal platform-specific quality patterns: MaFengWo exhibits prominent spatial centrality and stable temporal cointegration; TripAdvisor demonstrates simplified core-periphery structures with high volatility; Ctrip presents dynamic multicentricity particularly in Shanghai. Two domestic platforms, MaFengWo and Ctrip, expose systematic deficiency on the theme 'Decision-making Plan' (92.9 similar to 96.4% lacking operational suggestions), while international TripAdvisor emphasizes 'Practical Information' and 'Consumption Activity' but 40.76% neglects original viewpoints. Heterogeneous network analysis identifies the behavioral signatures of high-reliability user-preference attachment, quality stability, and profile homogeneity. This work bridges theoretical rigor with operational scalability, demonstrating the potential of LLMs in content governance for digital tourism.

Unexpected port disruptions, caused by severe congestion or geopolitical conflicts, can lead to reassignment of shipping flows and trigger cascading hazards across the Global Container Shipping Network (GCSN), with farreaching socioeconomic impacts. Understanding the mechanisms of such cascading risk propagation requires accurate simulation of dynamic flow reassignment and the development of a tailored diffusion model for maritime scenarios. This research introduces a novel cascading modeling framework that captures port-level dynamic shipping flow reassignment and assesses cascading failure diffusion in the GCSN. Using massive container vessel trajectory data, we design a directed and weighted GCSN, identify alternative ports for disrupted nodes through a combination of network topology analysis and geographic nearest-neighbor search, and refine port selection with a nonlinear trajectory cost function incorporating maritime distance, vessel load differences, and port size. Cascading failure is simulated by integrating the Motter-Lai load-capacity model with the dynamic reassignment process, and three maritime-specific vulnerability indicators are proposed to evaluate network resilience. The findings show that limiting port loads to 60 % of capacity or increasing port capacity by 40 to 60 % significantly mitigates cascading risk. It also appears that the European port system is found to be more susceptible to cascading disruptions than its East Asian counterpart. The proposed approach provides actionable insights and risk mitigation strategies for enhancing the resilience of global maritime logistics in the face of unexpected disruptions.

The transport security of liquefied natural gas (LNG) is a major challenge in global energy supply chains. Understanding the impact of port disruptions and strengthening the system's resilience are critical to ensuring global energy trade. This paper introduces a comprehensive framework for assessing the resilience of the LNG transport network. First, we construct a directed and weighted network model of the LNG transport system, and apply multidimensional centrality measures to evaluate the importance of individual ports. Next, we assess changes in the network structure caused by intentional port node disruptions, employing a Bow-tie network structure analysis. Additionally, we introduce a novel weighted network efficiency metric that incorporates link directionality and weight to assess the static resilience of the network. Furthermore, we develop an SIR (Susceptible, Infected, Recovered) transmission model that integrates relative weighted influence metrics for each port to showcase the dynamic spread of risk following a disruption at a port node. Finally, a dynamic resilience assessment framework leverages several key metrics to enhance a better understanding of the resilience of critical LNG transit ports. Our modeling methodology and evaluation framework offer a theoretical foundation for stakeholders to mitigate unexpected risks and safeguard against the diffusion of disruption risk.

To mitigate the climate change caused by carbon emission issues, zero-emission vehicles (ZEVs) and zero-energy buildings (ZEBs) have attracted increasing attention due to the significant proportion of energy-related carbon emissions from the transportation and building sectors. The energy-matching problem of ZEBs between demand and generation is widely noticed by academia, and the energy-sharing method using electric vehicles (EVs) has proved to be an effective approach to improve energy-matching performance. However, the stability issue of the grid-interactive performance caused by the unstable renewable energy generation and the negative impact on the road traffic of ZEVs for energy sharing receive limited attention. This paper proposes instantaneous and predictive control methods for a zero-emission system consisting of two zero-energy buildings using ZEV energy sharing to enhance the building-grid interaction stability and reduce the negative impact of ZEVs on road congestion. A genetic algorithm model is implemented in predictive control. The impacts of different ocean renewable energy types on energy matching, grid stability, and economic benefits are investigated. The results show that the instantaneous control can provide up to 71.0 % better grid-interaction stability performance than basic control. An average of 9.2 % enhancement in the stability performance can be further achieved after implementing genetic predictive control. When considering road impact in predictive control, the annual practical road impact changes from around-0.16 to around 0.13 to 0.15 under different scenarios, while the grid-interaction stability performance remains almost the same with the genetic predictive control that just considers grid stability.

Driver profiling can provide a human-centered approach to portraying individual travel behavior and revealing their motivation, objectives, and needs, thereby contributing to driving safety analysis, location-based service, and intelligent transportation. However, existing trajectory-based methods are limited to measuring low-level features, such as average speed and radius of gyration. Although these features can characterize specific observable behaviors, such as driving operation and movement range, they fail to depict stable traits underlying individual travel behavior. In this study, inspired by the Big Five Personality Traits, we model the driver profile through four fundamental trajectory traits: extroversion, openness, neuroticism, and conscientiousness, and quantify these traits by developing a Trajectory Trait Scale (TTS). Experiments on more than one million trajectories from 2,051 anonymized private vehicle volunteers over eight months demonstrate that our method can provide a valid representation of individual drivers’ mobility patterns and driving behaviors. Specifically, we validate the consistency between trajectory traits and vehicle customer service records of drivers, including life rescue, navigation service, violation query, and fatigue companion. Besides, we find that trajectory integrity, seasonal changes, and traffic conditions exert small but non-negligible impacts on the stability of trajectory traits. These findings can enhance the understanding of human behavior in various spatiotemporal contexts, and illuminate the relations between trajectory traits and personality traits.

Human mobility modeling and prediction are central research topics in GIScience. Although deep learning has led to significant advances in these fields, existing trajectory prediction models still face challenges in capturing the complexity of individual mobility behavior. Regression-based models often overestimate the diversity of human mobility, whereas classification models tend to underestimate it. This study attributes these biases to the models' limitations in recognizing the spatial relationships among activity locations and mobility heterogeneity across individuals. To address these challenges, we propose the Spatial Preference Map-based Transformer (SPM-Former), explicitly integrating spatial proximity and mobility heterogeneity to enhance trajectory sequence prediction. To capture individual mobility characteristics, SPM-Former utilizes the Spatial Preference Map (SPM) to represent individuals' spatial visitation preferences and adjacency relationships between locations. Then, we introduce two encoding modules to decode the information hidden within the SPM: one for encoding trajectory-level spatial-temporal information and another for embedding individual-level overall mobility features. Furthermore, we propose a novel optimization method, SPM-Loss, to assess prediction accuracy from the global spatial distribution perspective. Experimental results on a large-scale dataset from Japan demonstrate that SPM-Former outperforms state-of-the-art classification-based models, achieving approximately 3% and 20% improvements in trajectory sequence similarity and overall spatial feature similarity, respectively.

The digital economy drives economic growth and regional competitiveness. Understanding the evolution of county-level digital economies is essential for regional economic transformation, upgrading, and long-term development. Traditional assessment methodologies have several shortcomings for representing the county digital economy, especially data availability and reliability. In this paper, we develop a multi-scale analytic framework using complex network indicators including average clustering coefficient, k\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$k$$\end{document}-core, and weighted degree at macro, meso, and micro scales. The framework allows us to establish a county-level network using enterprise investment data from Fujian Province, China, to study digital economy investment development from 2000 to 2021. The outcomes are: 1) The digital economy's investment scale and connection in the county grew in stages, with network expansion aligning with the concept of"the rich leading the whole, and the whole leading the poor."2) The interconnectivity hot zones, which made up less than 9% of counties, had a major impact on the network and have gotten stronger. Investment linkage control increased from 24.64% in 2000 to 41.56% in 2021, and the focus of hot areas shifted from outside the province to within the province. 3) Over time, the top six key counties have increasingly controlled more than 30% of the total investment quota. In 2021, when 2% of counties controlled 60% of investment, developmental imbalances became more important.

With increasing concerns about overtourism and its associated impacts, there remains an imperative need to understand the spatial concentration of tourist activities within destinations. This study demonstrates that mobile phone data can be leveraged to accurately quantify the spatial concentration effect and potential heterogeneity among travelers with different social backgrounds. We introduce an analytical framework and demonstrate its efficacy through a case study of international travelers in three Korean cities (Seoul, Busan and Jeju). Results show that Asian travelers exhibit the highest spatial concentration, followed by European and North Americans, a finding that is consistent across all three cities and different times of the day. The Theil's entropy index effectively portrays spatial similarity of activities among the three groups, revealing that the spatial preferences of European travelers are intermediate between those of Asians and North Americans. The findings underscore the importance of considering travelers' social backgrounds in tourist flow management.

The complexity of global trade has led to intricate competition in the semiconductor manufacturing industry. Accurately assessing the trade pattern and competitive relations among countries is difficult when depending on a singular trade category or numerous isolated indicators. In this paper, directed weighted networks are built with the trade data for five categories of equipment from 2012-2022 and some geographically enhanced indicators are presented to investigate the network structure and resilience. Furthermore, a competitiveness assessment model using entropy-base is proposed to examine the spatial distribution and temporal changes of different countries' trade flows. The findings include: 1) Trade scale and 'small-world' characteristics have been strengthened, with 2018 as a turning point, marking an evolution from 'expansion' to 'consolidation'. Network resilience is significantly influenced by key nodes. 2) Global trade has become an oligopolistic structure, highly concentrated in eight key countries - Japan, Singapore, the USA, Germany, the Netherlands, China, the United Kingdom and France. 3) The national competitiveness exhibits a 'four-tier' structure. Regionally, Europe, Southeast Asia and East Asia are the most competitive. The Pacific, Africa and parts of South America offer significant potential for further market expansion.

Rheumatoid arthritis (RA) is a systemic autoimmune disease characterized by chronic inflammation of the synovial membrane, leading to cartilage destruction and bone erosion. Due to the complex pathogenesis of RA and the limitations of current therapies, increasing research attention has been directed towards novel strategies targeting fibroblast-like synoviocytes (FLS), which are key cellular components of the hyperplastic pannus. Recent studies have highlighted the pivotal role of FLS in the initiation and progression of RA, driven by their tumour-like transformation and the secretion of pro-inflammatory mediators, including cytokines, chemokines and matrix metalloproteinases. The aggressive phenotype of RA-FLS is marked by excessive proliferation, resistance to apoptosis, and enhanced migratory and invasive capacities. Consequently, FLS-targeted therapies represent a promising avenue for the development of next-generation RA treatments. The efficacy of such strategies – particularly those aimed at modulating FLS signalling pathways – has been demonstrated in both preclinical and clinical settings, underscoring their therapeutic potential. This review provides an updated overview of the pathogenic mechanisms and functional roles of FLS in RA, with a focus on critical signalling pathways under investigation, including Janus kinase/signal transducer and activator of transcription (JAK/STAT), mitogen-activated protein kinase (MAPK), nuclear factor kappa B (NF-κB), Notch and interleukin-1 receptor-associated kinase 4 (IRAK4). In addition, we discuss the emerging understanding of FLS-subset-specific contributions to immunometabolism and explore how computational biology is shaping novel targeted therapeutic strategies. A deeper understanding of the molecular and functional heterogeneity of FLS may pave the way for more effective and precise therapeutic interventions in RA.