Neo4j GDS for Snowflake
summary
Neo4j Graph Data Science runs inside Snowflake via the Snowpark Container Service so teams can run algorithms through SQL without moving data, using a pay‑per‑session model; projections read and write to Snowflake tables and results join with existing pipelines. The demo shows PageRank and community detection at scale — 2 billion nodes, 1 billion relationships, PageRank in ~3 minutes and writeback in ~12 minutes — and entity resolution that expands 241 labeled fraud accounts by 158 new risk candidates (≈+65%), all production‑ready.
AI-written summary · about
chapters (8)
- 0:00Introduction to GDS on Snowflake
- 2:49Graph data science use cases
- 5:45Technical architecture
- 8:03Implementing graph algorithms
- 9:14Deployment and performance
- 11:50Fraud detection demo: segmentation
- 20:19Fraud detection demo: entity resolution
- 25:42Resources and conclusion
related talks
26mGraph Analytics in Snowflake: Beyond Traditional SQL
29mFrom Warehouse to Wisdom: Activating Business Data with Neo4j
23mAcceleratinag Data Science and Machine Learning with Snowflake DataTech
13mLightning Talk: Highlights of the Snowflake Elastic Data Warehouse
29mBreaking Silos: Using Neo4j to Unify Tax and Finance Data
20mSnowflake concepts & architecture
29mScaling Python-Powered Machine Learning With Snowflake
23mDisrupting Customer Data Platforms: Neo4j vs. the Status Quo