2026-09-11

7 arXiv categories· 96 journals· 2541 candidates — 7 worth your time· 195 without an open abstract

Headline

An open-source workflow for extracting orientation-specific building envelope attributes from cadastral data

Today, together

tag shift
no tag ran above its 30-day average
shares references
no items share references today 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 109 of this day's 202 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
Environmental Earth Sciences; Institute of Archaeology of the Slovak Academy of Sciences; Instituto de Investigación Biomédica de A Coruña; and 5 more none of them on a paper we published in the 91 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.

The UrbanFormAnalyzer introduces an open-source Python library integrated with QGIS to automatically extract orientation-specific external perimeters and envelope descriptors from cadastral geometries. Another study produces annual 1-kilometer resolution global maps estimating the probability of 2G, 3G, and 4G mobile network coverage across 214 countries and territories from 1999 to 2030. Finally, archaeological excavations at Juukan Gorge recovered the first physical evidence of Tasmanian devils in northwestern Australia's Pilbara region, analyzing a Holocene dentary fragment and a pre-LGM Late Pleistocene premolar.

datapublished: Computational Urban Science

Daniel Garcia-Gonzalez, Jorge Rodríguez‐Álvarez, Miguel R. Luaces, et al.
Computational Urban Science

Universidade da Coruña

Urban Design and Spatial Analysis · 3D Modeling in Geospatial Applications · Remote Sensing and LiDAR Applications · land use data · Universidade da Coruña · Instituto de Investigación Biomédica de A Coruña

UrbanFormAnalyzer: an open-source workflow to derive orientation-specific building envelope attributes from cadastral data

The UrbanFormAnalyzer is an open-source Python library integrated with QGIS that automatically extracts orientation-specific external perimeters and envelope descriptors from cadastral geometries. It features a standardized input structure, includes an adapter to download and preprocess Spanish Cadastre KML files, and distinguishes exposed from adjacent walls by cardinal direction. The tool was tested by matching the generated spatial data with national energy certificates to estimate orientation-specific glazing ratios.

Why it matters — It automates the extraction of detailed building envelope attributes that are critical for urban energy modeling but are typically missing from national energy certificate databases.

Caveat: The predictive accuracy for estimating orientation-specific glazing ratios remains limited in the initial case study.

Computational Urban Science · pdf · doi

published: Archaeology in Oceania/Archæology & physical anthropology in Oceania

Timothy J. Churchill, Michael Slack, Elise Matheson, et al.
Archaeology in Oceania/Archæology & physical anthropology in Oceania

UNSW Sydney

Pleistocene-Era Hominins and Archaeology · Human-Animal Interaction Studies · Wildlife Ecology and Conservation · UNSW Sydney · Institute of Archaeology of the Slovak Academy of Sciences · Environmental Earth Sciences

First Record of Tasmanian Devils in the Pilbara: Late Quaternary Subfossils From Juukan Gorge, Puutu Kunti Kurrama Country, and Insights Into Human—Faunal Relationships and Extinction

Archaeological excavations at Juukan Gorge recovered the first physical evidence of Tasmanian devils in northwestern Australia's Pilbara region, consisting of a Holocene dentary fragment and a pre-LGM Late Pleistocene premolar. Analysis of these specimens and associated trace fossils, such as chewed and digested bones, documents a continuous human-devil interaction spanning at least 45,000 years until the species went extinct on the mainland approximately 3,200 years ago.

Why it matters — The findings establish that Tasmanian devils inhabited arid inland environments and maintained a long-term ecological relationship with humans, scavenging on butchery waste from the moment of human arrival until the introduction of dingoes disrupted this dynamic.

Caveat: The evidence of sustained interaction and ecological dependency is reconstructed from a single archaeological site complex.

Archaeology in Oceania/Archæology & physical anthropology in Oceania · doi

published: Journal of Housing and the Built Environment

Mateusz Tomal
Journal of Housing and the Built Environment

Krakow University of Economics

Housing Market and Economics · Spatial and Panel Data Analysis · Urban Planning and Valuation · spatial econometrics · Amsterdam · Krakow University of Economics

Econometric and machine learning spatially varying models for the mass appraisal of residential properties

The study compared the predictive performance of global and local econometric models against machine learning models for housing sales and rental prices in Warsaw and Amsterdam. It evaluated multiple regression, XGBoost, Geographically Weighted Regression (GWR), Multiscale GWR (MGWR), Geographically Weighted Random Forest (GWRF), Geographical XGBoost (G-XGBoost), and a hybrid GWRBoost model. GWRF and XGBoost achieved the highest out-of-sample accuracy, with GWRF significantly outperforming other models on sales prices, while both performed equivalently on rental prices.

Why it matters — It demonstrates that while modeling nonlinear relationships is sufficient for high predictive accuracy in rental markets, accounting for spatial heterogeneity is crucial for predicting sales prices. This provides automated valuation model developers with clear guidance on when to deploy computationally intensive local spatial models versus simpler global machine learning models.

Caveat: The findings are based on two European capital cities and may not generalize to housing markets with different structural dynamics or data availability.

Journal of Housing and the Built Environment · doi

datapreprint

Till Koebe, Theophilus Aidoo, Ali El Chami, et al.
arXiv cs.CY

A global mobile network coverage raster product at 1km resolution, 1999--2030

This study produces annual 1km resolution global maps estimating the probability of 2G, 3G, and 4G mobile network coverage across 214 countries and territories from 1999 to 2030. The dataset combines predictions from a machine-learning model (achieving an AUC of 0.89 to 0.92 on held-out countries), a techno-economic simulator, and a spatial deep-learning model, all trained on 2,409 operator-reported coverage maps up to 2020. Predictions for 2021 to 2024 rely on recent geospatial data, while 2025 to 2030 estimates are projected from demographic and infrastructure forecasts.

Why it matters — It provides the first globally consistent, sub-national, and longitudinal record of mobile connectivity, establishing a baseline to measure the digital divide and link network access to household-level socio-economic outcomes.

Caveat: Estimates for the years 2021 through 2030 rely entirely on model predictions and demographic projections rather than direct operator reports.

arXiv · pdf

preprint

Gul Sher Ali, Michail Giannakos, Monica Lillefjell, et al.
arXiv cs.CY

Digital Mental Health Interventions · Innovative Human-Technology Interaction · Urban Green Space and Health

Designing Technology for Social Wellbeing in Built Environments: A Conceptual Framework

This study synthesizes empirical literature to construct a conceptual framework for designing digital technologies that support social wellbeing in built environments. The framework consists of three core components: a model identifying specific social wellbeing dimensions for technology to target, a matrix classifying the types of contributions technology can make, and a mapping of the spatial or social levels at which these interventions operate.

Why it matters — It provides a structured, unified design guide for researchers, practitioners, and policymakers, establishing a conceptual link between urban digital interventions and preventative public health outcomes.

Caveat: The framework is conceptual and based on a synthesis of selected empirical studies rather than being validated through a new deployment or empirical test.

arXiv · pdf

preprint

Nathaniel Hendrix, Carl Y. Zhang, Chris Heitzig, et al.
arXiv stat.AP

gradient boosting · census data · United States

Geospatial Foundation Models Capture Health-Relevant Dimensions of Place Beyond Conventional Social Risk Indices

The study evaluated four geospatial foundation model families using 2022 satellite imagery to predict American Community Survey variables and 40 chronic disease and health-behavior outcomes from CDC PLACES across 82,646 census tracts in the contiguous United States. Using LightGBM, the models predicted housing type with an R-squared of up to 0.54, though they performed poorly on disability, unemployment, and income disparity. Crucially, the models explained up to 54% of the health outcome variance left unexplained by traditional social risk indices, with performance scaling from a mean explained variance of 0.31 in the smallest census tracts to 0.39 in the largest.

Why it matters — It demonstrates that satellite-derived geospatial foundation models capture physical characteristics of neighborhoods that directly relate to public health outcomes but are entirely missed by standard socioeconomic risk indices.

Caveat: The models' predictive performance is weaker for non-visual socioeconomic factors like unemployment, disability, and income inequality, and their accuracy decreases in smaller census tracts.

arXiv · pdf

preprint

Xue Qiu, Jianli Xiao
arXiv cs.LG

large language model

A Dynamic Fusion Large Language Model for Traffic Flow Prediction

The Dynamic Fusion Large Language Model (DF-LLM) was developed to improve traffic flow prediction by combining a spatiotemporal embedding module, a graph convolution-based fusion module, and an adapted LLM backbone. The model uses a differentiated parameter adaptation strategy, a context aggregation attention module, and residual connections to capture multi-scale spatiotemporal features and global dependencies. Evaluation across four benchmark datasets demonstrated that this architecture outperforms existing traffic prediction models.

Why it matters — It bridges the gap between traditional neural networks that rely on singular feature modeling and large language models that typically fail to capture spatial topology and dynamic spatiotemporal correlations.

arXiv · pdf

Still cited

Sherwin Rosen (1974), Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition 1 of today's items cite it · 21 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.