How Foursquare Built a Data Marketplace Using Metadata

About the Session

Foursquare’s new Geospatial Intelligence Platform revealed a key insight: many data scientists want to use location data in their ML workflows without needing deep geospatial expertise. To meet this need, Foursquare built the FSQ Spatial H3 Hub—a governed data marketplace powered by DataHub’s enterprise metadata capabilities.

Shirshanka and Vikram will discuss how metadata enables Foursquare to make geospatial datasets accessible, secure, and enterprise-ready, while supporting modern workflows across Spark, Python, and DuckDB.

Key Takeaways:

• How metadata management unlocks governed data marketplaces
• Ways to practically combine Iceberg catalogs with Spark, Python, and DuckDB
• Why lineage, versioning, and governance are critical for enterprise ML adoption
• Lessons from launching with 20+ open datasets to power machine learning at scale