From embeddings to spatial insights: Five new tools in CARTO Workflows

CARTO has expanded its Analytics on Embeddings Extension Package for CARTO Workflows with Embedding Profile, Mean Deviation, Neighborhood Deviation, Spatial Aggregation, and Vector Normalization, along with bivariate color mapping in its Visualization component. Designed to run directly in cloud data warehouses, these tools support analysis of geo-embeddings, which are vector representations of a location’s demographic, environmental, economic, or physical characteristics. Mean Deviation identifies anomalous or representative locations relative to global, group, or temporal averages, while Neighborhood Deviation measures differences between nearby spatial grid cells to reveal boundaries and local heterogeneity. Spatial Aggregation transfers embeddings from grids to business-relevant polygons such as ZIP codes, and Embedding Profile combines multiple vectors into weighted market or territory profiles for similarity searches, clustering, or change detection. Examples show the capabilities being used to assess differences within retail catchment areas, monitor changing urban patterns in Madrid using satellite embeddings, and identify New York insurance markets similar to high-performing California markets. Vector Normalization improves the efficiency of dot-product similarity calculations by converting vectors to unit length, where dot product corresponds to cosine similarity.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.