FHIR R4 Synthetic Data

**dbignite** is an open-source toolkit from Databricks designed to simplify working with healthcare data encoded in the FHIR standard. It enables extraction, flattening, and conversion of FHIR bundles into relational tables for analytics at scale.

**Why dbignite**

FHIR bundles are deeply nested JSON structures, which makes them difficult to query or analyze directly. dbignite flattens FHIR bundles into tabular form (Spark DataFrames or SQL tables), making them analytics-ready.
Once flattened and stored in a lakehouse, the data becomes suitable for large-scale querying, cohort analysis, machine learning workflows, and integration with other healthcare or non-healthcare datasets.
Built on PySpark and SQL, dbignite allows data teams to ingest, parse, and analyze FHIR data using familiar tools and frameworks.

**What dbignite Does**

- Reads FHIR bundles, supporting major FHIR versions (R4, R5, and “ci-build”), and allows selection of resource types.
- Parses and flattens nested FHIR data into Spark DataFrames or tables per resource (such as Patients, Claims, Conditions, or Encounters).

Synthetic FHIR data in R4 format. The notebook shows how to use an [open source Databricks solution accelerator ](https://github.com/databricks-industry-solutions/dbignite) to parse FHIR data into SQL tables to perform downstream analytics.

Sample data provided through [Synthea](https://synthea.mitre.org/downloads)