55K
Patient Records
3
Analytics Marts
dbt
Transformation
BigQuery
Warehouse
Production clinical data warehouse using dbt + BigQuery. Ingests 55,000 patient records through staging models into three analytics marts: condition KPIs, billing analysis, and monthly trend reporting. Full dbt docs, tests, and lineage graph included.
55K patient records, 3 analytics marts, full dbt lineage. The data warehouse pattern healthcare teams actually need.
Tech Stack
dbtBigQuerySQLPythonGCPdbt-testsJinjaYAML
Features
Staging Models
Raw clinical data → clean typed staging layer with dbt source tests and freshness checks.
Condition KPIs
Condition prevalence, readmission rates, LOS trends — analytics mart ready for BI tools.
Billing Analysis
Claim amounts, denial patterns, payer mix analysis across 55K patient encounters.
Monthly Trends
Time-series mart tracking month-over-month clinical and financial metrics.
How It Works
01
Source
Raw clinical CSVs → BigQuery raw dataset with dbt source definitions
→
02
Stage
dbt staging models: type casting, null handling, PII deidentification
→
03
Marts
3 analytics marts: condition KPIs, billing, monthly trends
→
04
Docs
dbt docs site with full lineage graph, test coverage, column descriptions