This fourth edition teaches Python-based GIS and geospatial analysis, covering data, mapping tools, remote sensing, elevation, modeling, and real-time processing. It guides readers toward building a geospatial application that can generate reports.
This textbook teaches readers to design and build agent-based models and connect them to GIS. It covers core concepts, QGIS, NetLogo examples, model testing, and spatial modelling of human behaviour.
Google Research outlines a Geospatial Reasoning agent that integrates Earth AI foundation models with planet-scale data to perform modular, explainable spatial analysis beyond traditional LLMs; BYOD support and cloud-native data access enable enterprise workflows, with early results showing improved health outcome predictions and higher spatial resolution.
Translated title. Original: Mercadini a Cervia: il viaggio tra le parole al Saretina 152 - AmeVe Blog
A data journalist claims to debunk geographic misinformation using satellite data and GPS, offering a rigorous, numbers-backed view of a broader reality and presenting a 'New World' unseen by others.
Translated title. Original: Libertas Livorno vs Elachem Vigevano: il test per la vetta in A2 - AmeVe Blog
A data-journalist explains their approach to debunk geographic misinformation using satellite and GPS data, emphasizing rigorous fact-checking to reveal a broader reality than mainstream narratives.
EarthRecord tests 12 multimodal agents on 238 land-change questions using four decades of raw satellite data from 56 sites. Agents often completed the analysis but struggled to tie images to real changes, and about half of correct answers shifted when satellite data were altered.
The document explains embeddings as numerical vectors in a high-dimensional latent space, highlighting the challenge for humans to interpret them and the ease for computers. It notes that geospatial data use may serve as a helpful guide, with crochet-themed illustrations to aid intuition.
Translated title. Original: ASCII.jp：インフラ点検・建設DXを加速する「Geo-Snap」、株式会社リコーとの販売連携を開始
Remodel Partners and Ricoh will jointly offer Geo-Snap with two Ricoh 360-degree cameras for construction, infrastructure, and local government work. Geo-Snap uses image location and direction data to map photos and create walkthrough VR tours; a Shiga Prefecture trial found it automated photo-to-map tasks that previously took 15 to 30 minutes per image.
The paper presents a framework to create spatial descriptors for predicting local spectral properties and hysteresis from images, balancing neighborhood size, spatial resolution, and descriptor dimensionality. It uses regression benchmarks, SHAP attribution, and a genetic algorithm to optimize region partitions and pooling, with Cartesian and polar representations; the method aims to reveal where detail is necessary and is applicable to molecules and solids with appropriate rules.
HOT discusses how GeoAI and community mapping support humanitarian response, highlighting data generation by locals, model fine-tuning for regions, and the need for human validation. It also covers the fAIr platform, STAC-MLM standardization, an open model-contribution call, and funding opportunities.