GeoGosha

https://www.alphaxiv.org/abs/2505.20148?chatId=019ce10e-f236-7274-997a-3a72993f5509 btw this was submitted to the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Track on Datasets and Benchmarks

кажется я понял у кого сейчас с деньгами хорошо (это был стадион «Краснодар»)

OSM Quality Checker https://www.gishub.io/sites/osm-quality-checker.html

🥴 мне снова забанили Claude

🍝🍝🍝 Person-based aggregate space-time accessibility (PASTA): Bridging the gap between place- and person-based accessibility https://linkinghub.elsevier.com/retrieve/pii/S0966692326000566

ха, а вот и вторая часть от них же (с их поддержкой) 🌍 In this work, we propose a reproducible workflow to enhance low-LOD 3D building models by reconstructing façade openings directly from widely accessible street view imagery (SVI). https://www.sciencedirect.com/science/article/abs/pii/S092658052600083X?via%3Dihub

кстати, недавно был у них в лабе, они оч классные и на самом деле идеологически близки к тому что мы делаем у нас в лабе только публикуются в nature

🔥 NUS did it again (сорри я просто их фанат) https://www.sciencedirect.com/science/article/pii/S0924271626000675?via%3Dihub In this work, leveraging the Urbanity and OpenFACADES packages, we developed a pipeline that integrates hierarchical urban features and cross-view imagery (remote sensing + street view) into a heterogeneous urban graph for building attribute prediction. Specifically:  (1) We construct a heterogeneous graph representation of urban environments, connecting buildings with multi-scale contextual elements to capture hierarchical structure 🏙️  (2) We introduce a feature propagation mechanism to compensate for missing street-level information, and propose a deep fusion module to integrate multi-modal signals 🏞️  (3) The framework is evaluated across three cities (Amsterdam, Berlin and Washington D.C.), predicting 10–12 building type categories from OSM—showing robust and consistent performance compared to prior approaches 🏠

Geoff Boeing: https://www.tandfonline.com/eprint/AV6BGDIHNEFIGXQPNPBW/full?target=10.1080/13658816.2026.2628193 free preprint: https://osf.io/preprints/socarxiv/qepc6_v1