2026-10-12

7 arXiv categories· 96 journals· 3270 candidates — 7 worth your time· 42 without an open abstract

Headline

A multimodal origin-destination flow benchmark for transit forecasting in Seoul

Today, together

tag shift
no tag ran above its 30-day average
shares references
Greening thy neighbor: The impact of cross-border high-speed rail network positions on transboundary air pollution in China shares 3 references with Innovation in the Periphery: Evaluating the Impact of High-Speed Rail on County-Level Innovation in China’s Yangtze River Delta Region (2026-07-20)
3 of them
  • Roles of accessibility, connectivity and spatial interdependence in realizing the economic impact of high-speed rail: Evidence from China
  • High-speed railways and collaborative innovation
  • The local effects and neighborhood effects of high-speed railway on urban entrepreneurial vitality: Evidence from China
; Greening thy neighbor: The impact of cross-border high-speed rail network positions on transboundary air pollution in China shares 3 references with High-Speed rail expansion and avian Biodiversity Decline: Evidence from a Hub-Link Perspective (2026-09-29)
3 of them
  • High-speed rail accessibility and haze pollution in China: A spatial econometrics perspective
  • High-speed rail and CO2 emissions in urban China: A spatial difference-in-differences approach
  • High-speed rail opening and urban air quality
counted over the 7 papers published above
on the same shoulders
no foundation work was cited by four of the day's papers counted over the 43 of this day's 49 named papers whose reference lists we have — standing on the same older work is not the same as being about the same thing
first appearance
Chung-Ang University; LIUPPA, UPPA; Pohang University of Science and Technology; and 1 more none of them on a paper we published in the 122 days before this issue

Two counts, not one counted twice: every row here but “on the same shoulders” counts the papers published above, and that one counts every paper this day named — those and the ones under Also published today, whose abstracts we could not read.

SeoulMMOD introduces an open benchmark dataset containing three years of hourly origin-destination flows across six transit modes in Seoul. A panel analysis of 626 Chinese provincial border counties reveals that improved cross-border high-speed rail network connectivity reduces PM2.5 levels. Additionally, spatial greening scenarios in Catania, Italy, were evaluated using thermal comfort simulations alongside dynamic human activity density estimates to measure climate regulation impact.

datapreprint

Taeyoung Yu, Seonbin Jo, Jiwon Kim, et al.
arXiv cs.LG

Traffic Prediction and Management Techniques · Transportation Planning and Optimization · Human Mobility and Location-Based Analysis · Seoul · The University of Queensland · Pohang University of Science and Technology

SeoulMMOD: A Large-Scale Multimodal Origin-Destination Flow Benchmark

This study introduces SeoulMMOD, an open benchmark dataset containing three years of hourly origin-destination flows in Seoul spanning six transit modes across 25 districts and 426 sub-districts. The authors evaluated 14 spatio-temporal forecasting baselines alongside contextual covariates including POI data, travel metrics, calendar attributes, and rainfall records. The benchmark revealed that several district-level models fail to scale to finer sub-district resolutions and that naive joint multi-mode training yields inconsistent accuracy improvements.

Why it matters — Researchers can now evaluate forecasting models on complete citywide travel patterns across multiple modes simultaneously, preventing mode-shift distortions that occur when analyzing isolated transit datasets.

arXiv · pdf

published: Applied Geography

Ding Li, Li Yue, Yuxuan Cao, et al.
Applied Geography

Lanzhou University

Energy, Environment, Economic Growth · Aviation Industry Analysis and Trends · Cross-Border Cooperation and Integration · network analysis · China · Lanzhou University

Greening thy neighbor: The impact of cross-border high-speed rail network positions on transboundary air pollution in China

An analysis of a panel of 626 Chinese provincial border counties from 2007 to 2020 demonstrates that improved cross-border high-speed rail (HSR) network connectivity reduces PM2.5 levels. Using a two-way fixed effects model, the study calculates that a one-standard-deviation rise in cross-border HSR degree, closeness, and betweenness lowers local PM2.5 concentrations by 0.72%, 0.75%, and 0.63%, respectively. The network integration is accompanied by stronger interregional trade links, reduced local government land allocations to polluting industries, and narrower pollution disparities between border counties and wealthier regions.

Why it matters — It demonstrates that expanding physical transport networks across administrative boundaries can directly mitigate transboundary pollution externalities in historically overlooked peripheral border zones.

Applied Geography · doi

published: Applied Geography

Mariacristina Sipala, Daniele La Rosa
Applied Geography

University of Catania

Urban Heat Island Mitigation · Urban Arborization and Environmental Studies · Urban Green Space and Health · land use data · University of Catania

For whom are these trees? Dynamic scenarios of urban greening for climate regulation in dense urban contexts with high resolution thermal comfort simulations

The authors developed and tested spatial greening scenarios in Catania, Italy, using the UMEP model to simulate microclimate cooling alongside dynamic GIS kernel density estimates of human activity. The analysis models socio-economic constraints to place new green infrastructure where thermal comfort improvements coincide with high densities of public space users across different times of day.

Why it matters — Instead of evaluating urban cooling solely through spatial area or canopy cover, this approach quantifies greening outcomes by the time-varying volume of people who actually experience reduced heat stress.

Caveat: The findings are derived from simulations in a single hot, dry Mediterranean city and rely on estimated user presence rather than empirical pedestrian counts.

Applied Geography · doi

preprint

Adriana Pompejano, Sébastien Laborie, Maxime Masson, et al.
arXiv cs.SI

LIUPPA, UPPA

Diverse Aspects of Tourism Research · Urbanism, Landscape, and Tourism Studies · Human Mobility and Location-Based Analysis

A Multi-Dimensional Score for Detecting Lesser-Known Points of Interest in Mountain Regions

The Obscurity Score framework synthesizes open spatial data across four dimensions—International Awareness, Attractiveness, Promotion, and Accessibility—to systematically identify and classify lesser-known points of interest in mountain environments. The authors built a generic data processing pipeline alongside a web-based prototype to operationalize and evaluate the scoring system.

Why it matters — By moving beyond single-variable metrics like social media counts or proprietary survey data, this approach gives destination planners an open-source tool to divert foot traffic away from congested alpine landmarks toward underutilized sites.

Caveat: The framework relies on preliminary results from a web prototype without providing empirical validation of actual visitor redistribution or long-term environmental impacts.

arXiv · pdf

preprint

Shuo Zhang, Aman Sunesh, Semiha Ergan, et al.
arXiv cs.RO

Trajectory Prediction for Autonomous Driving in Road Construction Zones Using Human-in-the-Loop Driving Data

Researchers created WZDrive-Data, a dataset of 1,183 work-zone driving segments collected from 33 participants using a human-in-the-loop VR simulator featuring worker animations. Utilizing driver gaze and control signals alongside scene contextual data, the authors developed WZDrive, a multimodal trajectory predictor achieving a minADE of 1.09 m and minFDE of 2.12 m at K=6. The approach outperforms leading baseline models on these metrics by 7.18% and 13.18%, respectively.

Why it matters — Integrating human driver gaze and control patterns into multimodal predictors improves autonomous vehicle trajectory forecasting through chaotic, temporarily reconfigured work zones where traditional map priors fail.

Caveat: Evaluation relies on simulated work-zone environments and virtual human interactions rather than real-world road construction data.

arXiv · pdf

preprint

Zhenhao Huang, Li Sun, Philip S. Yu
arXiv cs.AI

Data Management and Algorithms · Traffic Prediction and Management Techniques · Urban Design and Spatial Analysis

Looking Ahead and Back: Anticipatory Generative Modeling for Road-Network Trajectories

The study introduces Foresight, a graph-constrained generative model that frame trajectory generation on road networks as route planning through an 'anticipate-select-update-re-anticipate' loop. Instead of simulating plain location sequences, the method incorporates future route predictions toward a destination to guide deterministic transitions at complex intersections without drifting off the network. The framework outperformed existing diffusion and sequential baseline models in experimental evaluations.

Why it matters — This approach replaces pure spatial location generation with graph-aware transition modeling, eliminating off-network trajectory artifacts by anticipating downstream destination paths during step-by-step sequence generation.

arXiv · pdf

preprint

Anastasia Psarou, Mikolaj Rams, Blazej Torbus, et al.
arXiv cs.MA

street network data

Collective Autonomous Vehicle Routing Needs Social Awareness for Coexistence with Human Drivers

This study introduces a multi-agent reinforcement learning routing framework for autonomous vehicles (AVs) that incorporates marginal cost into the AV reward function based on counterfactual simulations without the AV present. Evaluated on road networks in Ingolstadt and Provins, social-aware routing with a 40% AV market share reduced the travel time penalty inflicted on human-driven vehicles from 1.7% under selfish routing to 1.1% in Ingolstadt, while simultaneously lowering AV travel times. A targeted approach that computes marginal costs for only the most impactful AVs achieved comparable network performance with substantially lower computational overhead.

Why it matters — It demonstrates that incorporating social awareness into AV routing algorithms resolves a coordination trap where selfish AV optimization hurts both human drivers and the AV fleet itself.

Caveat: The evaluated routing strategies rely on counterfactual traffic simulations to calculate marginal costs, which requires high computational resources and real-time network state estimates to deploy.

arXiv · pdf

Still cited

Carlos Moreno, Zaheer Allam (2021), Introducing the “15-Minute City”: Sustainability, Resilience and Place Identity in Future Post-Pandemic Cities 1 of today's items cite it · 62 of 3240 in the archive stand on it

Also published today

These appeared today in journals we track. Their abstracts are not openly available, so we cannot summarise them.