rw-book-cover

Metadata

Highlights

You’ve got three steps that stand between you and your finished curated dataset. If you don’t have an analytics engineer, then the work may be split up like this: (View Highlight)

This becomes even worse if the docs and validation are done by some other people

(View Highlight)

An analyst will have to use their experience to know when the dataset is “good enough” for the stakeholder and their question since 100% accuracy might not be the goal. And if we’re being honest, sometimes being directionally correct is all that’s needed to make a business decision. (View Highlight)

dbt makes it very quick to add data quality tests. In fact, it’s so quick, that it’ll take an analyst longer to write up what tests they want than it would take for an analyst to completely finish coding them (View Highlight)

we want to know why a certain logic was built into that specific model, then that’s where we’d turn to the documentation. (View Highlight)