Urban Currents

Shaping socio-functional urban identity: When morphometrics meet social media data to reveal neighbourhood structure and cores

The study presents a reproducible computational workflow that combines morphometric descriptors from MomePy-based tessellations with Foursquare-derived activity indexes. This integrated clustering framework was applied to 29 Southern European cities to map socio-functional urban structures and identify concentrated neighbourhood cores.

Why it matters — It demonstrates that combining spatial form with social media activity reveals functional differences within morphologically identical areas, as well as functional similarities across different physical forms, capturing a layer of urban organization that physical mapping alone misses.

Caveat: The functional activity patterns rely entirely on Foursquare data, which may bias the identified neighbourhood cores toward the demographic profile of that specific platform's user base.

Cities · doi

From physical surfaces to human-centric heat stress: LST and UTCI heat mapping reveals nonlinear effects of urban morphology

This study models and compares land surface temperature (LST) at 30-meter resolution and the universal thermal climate index (UTCI) at 1-meter resolution across Singapore. Using a geographically weighted XGBoost and generalized additive model workflow, it evaluates how 2D and 3D urban morphology factors influence both metrics, achieving test R-squared values of 0.855 for LST and 0.905 for UTCI.

Why it matters — It demonstrates that LST is an inadequate proxy for actual human heat stress because it fails to capture shading-driven and radiative processes, showing specifically that sky view factor dominates UTCI variability but has little effect on LST, and that higher albedo is unexpectedly associated with increased UTCI.

Caveat: The empirical findings and threshold relationships are derived from a single, highly dense tropical city-state, which may limit direct application to different climate zones or urban forms.

Sustainable Cities and Society · doi

Untangling the interplay between impervious surface landscape metric, thermal environment, and air pollution: a seasonal mediation analysis in Wuhan

This study constructed seasonal mediation models using remote sensing data from 2023 and 2024 in Wuhan, China, to analyze how impervious surface landscape metrics influence PM2.5 concentrations. It measures both the direct effects of these spatial patterns on air pollution and their indirect effects mediated through land surface temperature, while also accounting for the feedback loop of PM2.5 back onto surface temperatures.

Why it matters — It establishes that urban form metrics do not just correlate with air pollution directly, but actively alter local thermal environments which in turn amplify or mitigate PM2.5 levels, revealing a seasonal feedback loop that requires coordinated climate and pollution interventions.

Caveat: The empirical findings and seasonal mediation pathways are derived from a single case study of Wuhan, which may limit their direct applicability to cities with different climatic or morphological characteristics.

Urban Informatics · pdf · doi

Estimating and mapping Tokyo Window View Desirability using 3D model imagery and the hedonic price approach

This study developed a pipeline to estimate Housing Window View Desirability (HWVD) in Tokyo by combining photorealistic synthetic window-view images generated from Google 3D Tiles with a hedonic price model. Using rental records for three-bedroom properties, the model isolates the rent premium attributable to window views after controlling for structural and locational factors, which improved out-of-sample rent prediction R-squared from 0.62 to 0.71. Semantic analysis of the views revealed that higher desirability scores correlate positively with visible greenness and negatively with building views, with the highest scores clustered in wealthier southwestern wards like Shinagawa-ku and Minato-ku.

Why it matters — It establishes a scalable, automated method to quantify and map the economic value of residential window views across an entire metropolitan area, revealing a stark socio-economic disparity in access to high-quality visual landscapes.

Caveat: The study relies on synthetic window-view images generated from 3D models rather than actual photographs taken from inside the physical rental units.

Applied Geography · pdf · doi

Beyond scenic views: an integrative review of visual research on heritage landscapes—themes, methods, and typologies across scales

This review synthesizes 203 peer-reviewed studies published between 1995 and 2025 on visual heritage landscapes using thematic coding, VOSviewer keyword co-occurrence analysis, and exploratory cross-dimensional Sankey mapping. The analysis reveals that the literature is concentrated on analysis, evaluation, management, and conservation, relying heavily on perception-based empirical methods, GIS, and digital documentation of formally recognized, spatially bounded sites.

Why it matters — It establishes a structured organizing synthesis that maps research themes against landscape scales and methods, revealing critical gaps such as the weak coupling of perceptual and spatial evidence and the underrepresentation of informal, everyday, and hybrid landscapes.

Landscape and Urban Planning · doi

Who moves with the robotaxi? A role-based NLP sentiment analysis of acceptance, trust, and action in autonomous mobility

The study analyzed over 170,000 public comments regarding the Apollo Go robotaxi service in Wuhan, China, using a RoBERTa classifier and Sentence-BERT embeddings. It categorized public sentiment and discourse patterns based on three distinct situational roles: passengers, pedestrians, and motorists. The analysis revealed that while passengers generally report positive experiences, their routine use is hindered by operational boundaries and usability frictions, whereas pedestrians and motorists express significant concerns regarding safety, traffic unpredictability, and congestion.

Why it matters — It provides a scalable, digital-trace methodology to capture real-world user and non-user feedback under actual operating conditions, moving beyond traditional survey methods to pinpoint specific integration bottlenecks for autonomous vehicles in dense urban environments.

Caveat: The findings are based on social media comments from a single Chinese city, which may reflect localized regulatory conditions, platform demographics, and cultural attitudes toward autonomous driving.

Transportation Research Part A Policy and Practice · doi

Uncrewed aircraft system base site suitability analysis for multi-hazard disaster response

The study presents a multi-hazard geospatial and mathematical optimization framework to identify optimal pre-disaster launch bases for Uncrewed Aircraft Systems (UAS) in search-and-rescue operations. Tested in Japan's Mt. Unzen region, a territory vulnerable to landslides and tsunamis, the model integrates evacuation facilities, slope, land use, and seasonal hazard exposure data to map suitable zones. The analysis reveals that flexible, scenario-based base selection maximizes risk-weighted coverage, and that expanding drone flight ranges yields diminishing coverage returns beyond a specific threshold.

Why it matters — It provides disaster response planners with a systematic method to position drone launch sites under overlapping, multi-hazard threats rather than relying on single-hazard assumptions or ad-hoc deployment during emergencies.

Caveat: The framework's empirical findings are based on a single case study of the Mt. Unzen region in Japan, which may limit direct generalizability to regions with different topographical or hazard profiles.

Transportation Research Part D Transport and Environment · doi

Redefining empty trips through occupancy and transferability of cross-modal strategies: A review

This study reviews literature on empty trip generation across all transport modes and introduces a new continuous measurement framework based on the proportion of unused vehicle capacity rather than a binary empty-or-full classification. It also establishes a cross-modal framework to transfer mitigation strategies between different driver-provided mobility sectors, including public transit, carpooling, taxis, and ride-hailing.

Why it matters — By replacing binary definitions with a continuous capacity metric and mapping cross-modal similarities, the research enables planners to measure the true extent of underutilization and adapt successful reduction strategies from one transport sector to another.

Transportation Research Part D Transport and Environment · pdf · doi

Techno-economic analysis of HGV trailer tyres based on wear and fuel consumption

This study models the total cost of ownership for heavy goods vehicles by comparing conventional and low rolling resistance trailer tyres across long-haul, urban, and mixed-use driving cycles. The analysis integrates a tyre wear model with a vehicle energy consumption model to evaluate how trailer steering systems mitigate tyre scrubbing and affect overall costs.

Why it matters — It demonstrates that low rolling resistance tyres are only cost-effective in urban settings when paired with trailer steering systems, which can save up to £35,000 over a trailer's service life and cut mixed-use operational costs by 4.9%.

Caveat: The findings are based on simulated tyre wear and vehicle energy consumption models rather than empirical fleet tracking data.

Transportation Research Part D Transport and Environment · doi

Using system dynamics to explore shared mobility and sustainability connections

This study develops qualitative system dynamics models and causal loop diagrams to analyze the systemic impacts of shared micromobility, carsharing, ridesharing, and on-demand ride services on urban transportation. By synthesizing variables and causal relationships across multiple existing studies, the models map the non-linear interdependencies and feedback loops that characterize urban mobility systems.

Why it matters — It provides a holistic framework that visualizes the indirect, long-term consequences of shared mobility policies, demonstrating how interventions favoring a single transit mode can unintentionally undermine overall urban sustainability.

Caveat: The developed models are qualitative rather than quantitative, meaning they map the structure of system feedbacks without simulating numerical trajectories or specific local parameters.

Transportation Research Interdisciplinary Perspectives · doi

Predicting vessel arrival time in inland waterways: Integrating ETA, AIS, and traffic flow data

This study developed a vessel arrival time prediction framework for inland waterways by integrating port call records, AIS trajectories, and vessel characteristics. Tested on the Port of Rotterdam, the system reconstructs routes to calculate sailing distances and traffic flow, feeding these features into ensemble tree-based models. The model reduced the mean absolute error of arrival predictions by 79.55%, dropping from 17.06 hours in self-reported estimates to 3.49 hours.

Why it matters — It provides a highly accurate prediction method tailored to the structured routes and high traffic density of inland waterways, replacing unreliable self-reported vessel estimates with data-driven predictions that enable proactive berth scheduling.

Caveat: The framework's predictive performance was validated using data from a single port case study in Rotterdam.

Transportation Research Interdisciplinary Perspectives · doi

A reproducible data-driven framework for national-scale road-ferry accessibility in South Korea

This study developed a reproducible optimization framework to analyze South Korea's national road-ferry network using an 85 x 85 origin-destination matrix generated via the OpenRouteService API. It benchmarked multiple optimization algorithms, including Simulated Annealing and the exact Concorde solver, across two maritime routing scenarios: a multi-port setup and a single-port setup. The multi-port configuration reduced travel distance by approximately 17.2% and travel duration by 8.4% compared to relying on a single port.

Why it matters — It establishes a scalable, open-source workflow that demonstrates how choosing between distance-based and duration-based optimization criteria structurally alters routing topology, even when aggregate efficiency remains similar.

Transportation Research Interdisciplinary Perspectives · pdf · doi

Tree Value Visions – integrating relational values and multispecies justice in urban treescape management

This paper introduces 'Tree Value Visions', a framework designed to help local authorities integrate relational values and multispecies justice into urban forest management. It addresses the challenge of balancing competing policy objectives—such as carbon sequestration, biodiversity, flood regulation, and public health—under the spatial and resource constraints of urban areas.

Why it matters — It provides planners and developers with a structured approach to make evidence-led decisions on nature-based solutions, moving beyond simple tree-planting targets to consider broader ecological and social justice outcomes.

Caveat: The abstract outlines a conceptual framework and policy context rather than presenting empirical findings, case study evaluations, or quantitative performance metrics.

npj Urban Sustainability · pdf · doi

Publisher Correction: Anthropogenic Factors Dominate Taxonomic Diversity of Urban Street Trees While Climate Drives Phylogenetic Diversity

This is a publisher correction notice correcting the corresponding author designations for Yuandong Hu and Liangjun Da. No scientific content or data was altered.

Why it matters — The correction ensures proper academic attribution and contact information for the study's authors.

npj Urban Sustainability · pdf · doi

"We are all in big trouble! *Shock Emoji": Personal Narratives in Expressing Emotions, Opinions, and Data Regarding Climate Change in TikTok Short Videos

This study analyzed 200 TikTok videos tagged with #climatechange to categorize how creators communicate environmental issues, identifying four distinct content types: expression-feelings, views-appeals, news-information, and trend-hijacking. The analysis reveals that creators frequently use humor to deliver sharp critiques, substitute complex systemic debates with personal lifestyle narratives like vegetarianism, and reframe fragmented news media to project scientific authority while maintaining emotional resonance.

Why it matters — It demonstrates how short-form video platforms allow creators to translate complex scientific information into highly personalized, empathetic narratives, showing how these interactions help build community on a platform that lacks formal community structures.

Caveat: The findings are based on a relatively small sample of 200 videos under a single hashtag, which may not capture the full diversity of climate discourse on the platform.

arXiv · pdf · doi

Inter-Stop Energy Prediction and Causal Driver Quantification for Dual-Source Trolleybuses via a Time-Aware Tabular Deep Learning Architecture

A time-aware tabular deep learning framework was developed to predict inter-stop energy consumption for dual-source trolleybuses, integrating periodic time encoding with a batch-ensemble backbone. Tested on the Zurich trolleybus dataset combined with hourly weather records, the model achieved a Mean Absolute Percentage Error of 6.52% and an R-squared of 0.982, outperforming ten baseline models. A three-layer causal explanation pipeline identified regenerative braking ratio and average speed as the primary energy-saving factors, while coasting distance was the main driver of excess energy use.

Why it matters — The model moves beyond black-box prediction to quantify the actual causal drivers of trolleybus energy consumption, establishing actionable thresholds for driving behavior, vehicle technology, and catenary network planning.

Caveat: The empirical findings and causal thresholds are based on a single transit system in Zurich and may not directly generalize to cities with different topographies, climates, or fleet configurations.

arXiv · pdf