Instantly enrich your data science projects with H3-Indexed spatial features
H3 indexing engine
Our proprietary engine transforms otherwise inaccessible geospatial data — raster and vector — into ready-to-use H3 features at scale, no specialized processing required. Turn messy, fragmented sources into clean, analysis-ready features with a single step.
Universal access through Iceberg
One connection string. Our iceberg catalog works with your existing stack – Spark, DuckDB, Python – giving every data scientist instant access to production-ready geospatial features. Join spatial features with a simple SQL query – no GIS expertise required.
Production ready data catalog
20+ pre-indexed geospatial datasets — from global points of interest to census data, infrastructure, and environmental features — all standardized at H3 resolution 8 and available through an Iceberg catalog that works seamlessly with your existing data stack. Built with DataHub metadata management system to ensure a future-proof foundation that facilitates versioning, lineage tracking, and governance.
Coming soon
Available now in preview for free, with continued use pricing coming soon.
New open datasets added every month
Stay ahead with a constantly expanding library of H3-indexed features. We add fresh, open datasets regularly to power even more use cases.
Bring & publish your own proprietary datasets
Soon, you’ll be able to index and publish your own private geospatial data through our streamlined pipeline and keep everything discoverable in one place.
REST interface for real-time ML inference
Integrate spatial features directly into your production workflows with a robust REST API — ideal for streaming and real-time machine learning.
Explore our production-ready data catalog
FSQ OS Places
Global points of interest
Census & Population
Census block groups, tracts, and global population
Nature & Weather
Fire perimeters, fire severity, and water
Infrastructure
Roads, mineral and cell towers
Get started with Foursquare Spatial H3 Hub
Transform messy and scattered data into clean, H3-indexed tables.