🏥 55.5K
Encounters modeled
🧱 TMDL
Power BI as code
🔌 11
API routes
📊 Fabric
Lakehouse proof
Tech Stack
Microsoft FabricPower BIdbtFastAPIPythonTMDLDAXMLflowAzure AD
Before / After
Before
Before
Dashboard-only BIData
↓
Manual Power BI
↓
Pretty chart
↓
No lineage
×
Measures live inside a .pbix binary nobody can review.
×
No tests run before the data reaches a chart.
×
No API layer — the numbers can't be reused downstream.
×
'Trust me it's right' instead of 'here's the file that proves it'.
After
After
Git-reviewed Fabric stackdbt star schema
↓
TMDL in git
↓
Fabric validate
↓
Traceable BI
✓
BI as CodeCertified Power BI measures + relationships live in git as TMDL — reviewed like application code, not buried in a binary.
✓
Tested Before Prettydbt star schema with clinical data-quality tests runs before anything reaches a chart.
✓
Fabric ReceiptsLakehouse + semantic-model validation captured as markdown proof — not a screenshot you have to trust.
✓
Honest AUCXGBoost baseline logged in MLflow with real metrics, honest about the limits of synthetic demo data.