From Data Silos to Patient 360: Building Real-Time Cohort Analytics With Zero-ETL Graph Technology

Healthcare organizations generate 50+ petabytes annually, yet 80% remains siloed. This session demonstrates how PuppyGraph's zero-ETL graph layer transforms fragmented Databricks lakehouse data into unified patient 360 views without moving a single byte.The challenge: Identifying high-risk patients requires complex 50+ line SQL queries—slow and inaccessible.The solution: Graph technology excels at entity resolution. By linking attributes (addresses, providers) into networks, it matches records like "John A. Smith" and "J. Smith" through relationship patterns. We’ll compare traditional SQL joins against five-line graph traversals following the patient journey: Patient → Condition → Visit → Prescription.Graph queries mirror clinical thinking—connected journeys, not isolated tables.Key takeaways:

• Zero-ETL graph analytics on lakehouse data
• Integrating AI agent-ready graph infrastructure
• Unified governance through Unity Catalog