R Training Resources
R is the historically preferred language of the OHDSI community, and much of the ecosystem's ready-made tooling is built in R.
HADES — OHDSI's R Package Ecosystem
HADES (Health Analytics Data-to-Evidence Suite) is a collection of open-source R packages for large-scale analytics against the OMOP CDM. Key packages include:
| Package | Purpose |
|---|---|
| DatabaseConnector | Connect to OMOP databases (Redshift, PostgreSQL, etc.) |
| SqlRender | Write SQL once, translate to any dialect |
| CohortGenerator | Generate cohorts from ATLAS definitions |
| FeatureExtraction | Extract patient-level features for modeling |
| PatientLevelPrediction | Build and evaluate predictive models |
| CohortDiagnostics | Evaluate cohort definitions across databases |
Connecting to Emory's OMOP Data Lake from R
library(DatabaseConnector)
connectionDetails <- createConnectionDetails(
dbms = "redshift",
server = "<host from email>/<database from email>",
port = 5439,
user = "<your username>",
password = "<your password>"
)
conn <- connect(connectionDetails)
# Query the person table
querySql(conn, "SELECT COUNT(*) AS person_count FROM cdm.person")
disconnect(conn)
Use DatabaseConnector over RPostgres
While you can connect with RPostgres or DBI directly, DatabaseConnector is recommended for OHDSI workflows because HADES packages expect a DatabaseConnector connection object.