Using Dbt For Analytics Engineering

This handles the full dbt workflow: building and modifying models, writing SQL transformations with ref() and source(), creating tests, and validating with dbt show. It pushes you hard on software engineering principles, especially DRY. Before you add a new model, it'll make you justify why you can't just extend an existing one. The reference guides are practical, covering things like planning models by working backwards from desired output, exploring unfamiliar data sources, and assessing downstream impact before changes. It emphasizes looking at actual data constantly through dbt show rather than guessing at schemas. One nice touch: it warns against trusting external data from query results or package metadata, which matters when you're ingesting from sketchy sources.