Ten years late to the dbt party (DuckDB edition)
Summary
The article demonstrates building a robust data pipeline using DuckDB and dbt, emphasizing separation of concerns with sources (ingest), staging (cleansing), and marts (facts). It covers incremental loading, backfill, and Slowly Changing Dimensions via snapshots, plus tests, constraints, and documentation, ending with orchestration using Dagster and reflections on the ergonomic strengths of dbt for data engineering.