Important
This is a simple script to generate core OMOP tables with minimum levels of synthetic (fake) data, without any view towards realism. It is designed to support development of query and visualisation functionality, not for analysis. It is only populated with a small number of concept types (hardcoded in src/datagen.py) which are generated with a minimum level of clinical realism.
To use:
-
Configure cohort size and number of visits in
create_synthetic_omop.py -
Generate data:
python create_synthetic_omop.py -
Validate OMOP constraints:
duckdb -init validate_omop.sql. You will need to install the duckdb cli for this. If OMOP compatible, you should receive row counts for each table. Note that this does not yet validate against the vocabulary table.