INTERGEO Agenda - english

Geospatial data is not only about coordinates. With LiDAR, remote sensing, sensor networks, digital twins, IoT, and AI-based processing pipelines, geographic objects are increasingly described by rich sets of attributes: height, shape, condition, context, similarity, uncertainty to name a few. These attributes form high-dimensional feature spaces — the same conceptual space in which many AI models represent and compare objects. This talk explores how such feature spaces can become accessible for geospatial analysis. Instead of looking only at where objects are located on a map, we ask: Which objects are similar? Which ones are unusual? Which spatial patterns only become visible when thematic, structural, and analytical attributes are considered together? Using examples from multivariate geospatial data, including urban tree inventories derived from mobile mapping and LiDAR, the talk introduces an interactive visual analytics perspective in which geospatial objects can be rearranged according to selected features, similarity relations, and spatial criteria. The aim is not to replace maps, but to complement them: by moving between geographic space and feature space, analysts can discover latent groups, outliers, correlations, and AI-relevant patterns that remain hidden in conventional map views. The talk invites GIS professionals, researchers, and decision-makers to rethink geospatial visualization as an interface between maps, data science, and AI-assisted spatial understanding.