Urban Currents

GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation

The GeoFlow framework models origin-destination (OD) flows by integrating geospatial attributes, such as relative positions, k-hop, and geodesic distances, into area representations. It combines a geometric-intrinsic fusion encoder with an axial-global attention decoder, and pairs with flow matching models to generate synthetic mobility samples.

Why it matters — It overcomes the limitations of traditional graph-based methods that ignore global spatial structures, enabling more accurate predictions of long-range, multi-area dependencies and producing highly realistic, diverse mobility patterns.

arXiv · pdf · code

VOID URBANISM: Unbuilt New Cities and State Formation in the Democratic Republic of Congo

This paper analyzes the planning process of Kitoko City, a new city project in the Democratic Republic of Congo launched in 2019 that was never physically constructed. It introduces the concept of 'void urbanism' to examine how unbuilt, planned cities function as strategic tools of government that trigger land speculation, alter institutional practices, and shape national narratives of modernity.

Why it matters — It demonstrates that unbuilt urban projects are not merely financial or technical failures, but active political fictions that consolidate state power and organize social expectations through anticipation and latency.

Caveat: The study is based on a single qualitative case study of a planned city in the Democratic Republic of Congo, which may limit its direct applicability to different political and institutional contexts.

International Journal of Urban and Regional Research · doi

GROWING GREEN IN US SUBURBIA: Property Value, Race and Domestic Landscaping Aesthetics in Mid‐Century America

This historical case study analyzes how mid-century American redlining and racial exclusion of Black homebuyers were actively supported by a standardized domestic landscaping aesthetic. Promoted by banks, developers, and the federal government to stabilize property values, this green aesthetic functioned as a visual signal of racial exclusion and actively shaped a white suburban identity motivated to maintain segregation.

Why it matters — It challenges the prevailing assumption that urban greening is inherently positive and that environmental injustice is merely a problem of unequal spatial distribution, demonstrating instead how specific natural aesthetics can be engineered to enforce racial and economic segregation.

Caveat: The study relies on a historical case study of mid-century US suburbanization, which may not directly reflect the dynamics of contemporary urban greening initiatives.

International Journal of Urban and Regional Research · doi

Geographical dispersion and transmission of multiple SARS-CoV-2 variants in Hong Kong

The study integrated 1,725 viral genomes linked to patient residential addresses with urban mobility data to reconstruct the spatiotemporal spread of SARS-CoV-2 across epidemic waves three to five in Hong Kong. Using Bayesian phylodynamics, it identified that these waves originated in the high-density districts of Wong Tai Sin, Yau Tsim Mong, and Kwai Tsing, which then acted as the primary exporters of regional transmission. Statistical modeling revealed that population density and Mass Transit Railway (MTR) usage were consistent predictors of transmission rates between districts.

Why it matters — It demonstrates that combining pathogen genomic surveillance with intra-city transit and demographic data can pinpoint specific neighborhood-level transmission hotspots and pathways, providing a framework for targeted public health interventions in high-density urban environments.

Urban Informatics · pdf · doi

Assessing the potential of Tehran's green space master plan to enhance urban ecological performance, cooling capacity, and spatial equity

This study models the ecological and thermal impacts of Tehran's Green Space Master Plan (GSMP) using Biotope Area Factor (BAF) and Urban Cooling Capacity (UCC) indices. The proposed plan increases the city's overall BAF from 0.11 to 0.20 and its UCC from 0.146 to 0.211, though Gini coefficient analyses reveal these benefits disproportionately favor affluent northern districts. To address this spatial inequity, a modified scenario integrating green roofs in dense, disadvantaged neighborhoods was simulated, raising their local BAF from 0.06 to 0.35 and UCC from 0.18 to 0.35.

Why it matters — It quantifies the spatial inequities inherent in a major city's official greening plan and demonstrates how targeted vertical greening interventions can retroactively balance ecological benefits in dense, low-income areas.

Caveat: The evaluation relies on simulated index improvements rather than empirical, post-implementation measurements of temperature reduction or ecological health.

Applied Geography · pdf · doi

Health, built environmental, and environmental perceptions in relation to non-commute walking among middle-aged adults

This study analyzed survey data from 366 adults aged 45 to 65 in Rotterdam and Maastricht, the Netherlands, to examine how BMI, self-perceived health, and environmental features relate to non-commute walking frequency. Using zero-inflated Poisson regression models, the researchers measured objective environmental characteristics within 300 m, 600 m, and 1000 m buffers of homes alongside self-reported environmental perceptions. The analysis revealed that BMI, street connectivity, and perceived aesthetics were negatively associated with walking frequency, while address density and perceived safety showed positive associations.

Why it matters — The study demonstrates that objective and perceived environmental features contribute equally to explaining non-commute walking behaviors, and that these environmental factors influence walking independently of an individual's health status rather than being moderated by it.

Caveat: The findings rely on self-reported walking frequencies and environmental perceptions, which may introduce recall bias, and the sample is limited to middle-aged adults in two Dutch cities.

Cities · pdf · doi

Spatial inequalities in urban infrastructure for physical activity in Greater Santiago, Chile: Analysis by municipality and neighborhood

The study evaluated 23 indicators of physical activity infrastructure across 34 municipalities and 1,623 neighborhoods in Greater Santiago, Chile, grouping them into four dimensions: active transport, green areas, public transport, and safety. Green areas were identified as the most unequally distributed resource, while public transport was the most evenly distributed, even favoring lower-income municipalities. Except for public transport, the variation in infrastructure quality was greater within individual municipalities than between different ones.

Why it matters — It demonstrates that urban infrastructure disparities are more pronounced at the neighborhood scale than at the municipal scale, proving that municipal-level analyses mask significant localized inequalities and that urban interventions must target smaller geographic units.

Caveat: The study relies on an unweighted mean of indicators to construct its summary index, which assumes each of the 23 indicators contributes equally to physical activity promotion.

Cities · pdf · doi

Binocular visual field loss and motor vehicle crash frequency and severity: A generalized structural equation modeling study using linked data

This study analyzed linked ophthalmological, crash, and health records from 1990 to 2019 in Western Australia to evaluate how binocular visual field loss (VFL) affects driving safety. Using a generalized structural equation model on 4,307 drivers aged 50 and older involved in 5,537 crashes, the researchers found that binocular VFL directly increased crash frequency by 3% but had no direct or indirect impact on crash severity. Additionally, older age (70+) indirectly increased crash frequency through its association with VFL, while female sex and previous cataract surgery were linked to fewer crashes.

Why it matters — The research establishes a direct, quantified link between objective visual field deficits and crash frequency while demonstrating that crash severity is driven by crash type rather than visual impairment. This provides empirical evidence to support the inclusion of objective visual field testing in fitness-to-drive assessments for older adults.

Caveat: The study population is restricted to drivers who were already involved in at least one crash, which may limit the generalizability of the findings to the broader driving population.

Journal of Transport & Health · pdf · doi

Mapping urban environmental stress in Barcelona: a multivariate approach using GWPCA

The study constructs a multidimensional Environmental Stress Index for Barcelona by applying Geographically Weighted Principal Component Analysis to six environmental variables across 1,068 census tracts. The analysis reveals that 45% of the city's population resides in high-stress areas, which are concentrated in central, eastern, and northern districts covering just 25% of the municipal territory.

Why it matters — It replaces uniform, city-wide environmental assessments with a localized weighting system that identifies which specific stressors dominate in different neighborhoods, allowing planners to target interventions to distinct local inequalities.

Sustainable Cities and Society · doi

Data-Driven Predict-then-Optimize framework for landing scheduling with hybrid descent modes

This paper develops an integrated arrival management framework that combines a Kolmogorov-Arnold Network for predicting boundary estimated times of arrival with a mixed-integer linear programming model to jointly optimize arrival sequencing and descent-mode assignment. Validated using real-world operational data from Hong Kong International Airport, the framework incorporates the combined effects of aircraft types and descent-mode pairings directly into minimum separation constraints. The approach reduces total fuel consumption by 4.4% compared to actual operational baselines.

Why it matters — It demonstrates that universally maximizing Continuous Descent Operations is counterproductive under high-density traffic, as the resulting scheduling rigidity induces excessive delays that outweigh the fuel-saving benefits of the descent profile.

Caveat: The framework's performance and fuel-saving benefits are demonstrated using data from a single major aviation hub, Hong Kong International Airport.

Transportation Research Part B Methodological · doi

A bicycle traffic prediction framework using pretrained large language models

The researchers developed BiSTLLM, a spatial-temporal prediction framework that adapts pre-trained large language models to forecast short-term bicycle traffic. The model represents historical traffic sequences as tokens, integrates a causal temporal convolution processor to capture delayed weather impacts, and incorporates semantic embeddings of local POIs and bike lane accessibility. Evaluated on real-world bicycle flow datasets, the framework outperformed existing state-of-the-art models, maintaining high accuracy even in few-shot scenarios with limited data.

Why it matters — It demonstrates that general-purpose language models can be successfully fine-tuned for specialized urban mobility tasks, offering a viable forecasting method for cities with sparse sensor networks or limited historical cycling data.

Caveat: The abstract does not disclose the specific cities, dataset sizes, or baseline models used to validate the framework's performance.

Transportation Research Part C Emerging Technologies · doi

Assessing Governance Alignment for Shared and Micromobility: Evidence from Europe and the GCC

This study compared governance priorities and key performance indicators (KPIs) for shared and micromobility between Europe and the Gulf Cooperation Council (GCC) using a modified Delphi Poll, surveys, and semi-structured interviews with regional mobility experts. Survey data was analyzed using the Analytic Hierarchy Process (AHP) alongside reflexive thematic analysis of the interviews. The results show strong alignment on operational priorities like safety, reliability, and accessibility, but reveal that differences in regulatory maturity and institutional capacity alter how these KPIs are implemented in practice.

Why it matters — It demonstrates that while strategic-level governance frameworks for shared mobility can be adapted across highly different regional contexts, the actual application of regulatory oversight requires localized adjustment rather than direct policy copying.

Caveat: The findings rely on qualitative interviews and subjective expert rankings rather than empirical measurements of policy outcomes or operational data.

Journal of Urban Mobility · pdf · doi

Author Correction: Urban congestion relief experiments through routing-app interventions

This correction amends incorrect highway route numbers for Atlanta in the text of a previously published study on routing-app interventions and urban congestion relief.

Why it matters — The correction ensures the geographical accuracy of the highway routes referenced in the paper's analysis of traffic dispersion mechanisms.

Nature Cities · pdf · doi

The politics of green in Leinefelde’s housing environment

This study traces the spatial and functional evolution of green spaces in the social housing estates of Leinefelde, Germany, across two distinct political eras separated by the Wende. Using an interpretive historical analysis of GDR-era primary sources, including the journal Deutsche Architektur, the research documents how a socialist collective landscape was restructured into a fragmented system of green spaces through 1990s master planning and 2000s housing-scale interventions in the Plattenbau fabric.

Why it matters — It demonstrates how urban green space is not a politically neutral element but a dynamic medium that materializes shifting ideological discourses, showing how state-led welfare policies and post-reunification restructuring physically redefined the boundaries between public and private dwelling spaces.

Caveat: The empirical evidence is drawn from a historical case study of a single small town in former East Germany, which may limit direct generalizability to other post-socialist urban contexts.

Planning Perspectives · doi

Infinite Drive: Optimal Urban Location of Dynamic Wireless Charging at Signalized Intersections

The study develops a mixed-integer programming model to optimize the placement of Dynamic Wireless Power Transfer (DWPT) infrastructure at signalized intersections, accounting for vehicle acceleration, deceleration, and queue-dependent dwell times. Applying this model to Kawagoe City, Japan, and running Monte Carlo simulations of continuous trip chains up to 800 km, the researchers found that electrifying just 1.5% to 2.9% of the road network is sufficient to keep electric vehicles with 12 kWh batteries running continuously without plug-in charging.

Why it matters — It demonstrates that strategic, intersection-focused wireless charging can eliminate the need for large vehicle batteries, avoiding an estimated 1.7 to 3.0 tonnes of CO2 equivalent emissions from battery manufacturing per vehicle compared to a standard 40 kWh urban electric vehicle.

Caveat: The findings are based on a simulation model applied to a single Japanese city and rely on probabilistic signal patterns and saturation headway assumptions that may vary in other urban contexts.

arXiv · pdf