# Snowflake reference path

This path reproduces the core analytical model while allowing Power BI or another approved consumption layer to remain in place.

1. Land source batches into raw schemas using the client's approved ingestion tools and storage boundary.
2. Preserve source metadata and load history before standardized and curated transformations.
3. Build conformed location, date, payer, role, and service-line dimensions with facts at their declared grains.
4. Separate ingestion, transformation, serving, and administrative workloads with right-sized virtual warehouses and resource monitors.
5. Apply role-based access, masking or row-access policies where required, network controls, and audited administrative access.
6. Serve governed models to Power BI or another client-approved analytical layer.
7. Monitor query behavior, credit use, pipeline freshness, quality failures, and reconciliation status.

Production design must test workload isolation, data sharing, recovery, region, retention, cost, support ownership, and the semantic access pattern.