Tutorials

This section contains Jupyter notebooks that teach the cdmconnector package through a story-driven path. Each notebook follows: Setup → Explore → Build → Interpret → Exercises → What we learned.

Recommended order

#	Notebook	Topic
00	Welcome: Your first CDM	Connect, list tables, person count, preview
01	Lazy queries and pipelines	filter/select/join, compile SQL
02	OMOP people, time, domains	Demographics, observation period, visits
03	Story: GiBleed end-to-end	Cohort, index date, windowed analysis
04	Measurement distributions	Numeric stats, missingness, guardrails
05	Cohorts 101	Tables, attrition, counts
10	Pedagogical OMOP (GiBleed)	Mini-course: cohort, baseline, time-at-risk, outcome rates

Running locally

1. Environment: Create a venv and install the package with DuckDB: pip install -e ".[duckdb]". If you want the GitHub-install equivalent of R’s devtools::install_github(), use pip install "cdmconnector @ git+https://github.com/OHDSI/pyCDMConnector.git".
2. Eunomia: Example data is downloaded on first use via cc.eunomia_dir("GiBleed"), or run cc.download_eunomia_data("GiBleed") in advance.
3. Paths: Run Jupyter from the repository root so docs-src/assets/cohort_json and other paths resolve.

See Contributing for full setup; see Development for rebuilding the docs after editing notebooks.