Understanding Data Quality

Accuracy, completeness, consistency, timeliness, and fitness-for-use are common lenses. A dataset can be internally consistent yet wrong for your question if definitions shifted—see metadata and provenance.

Profile key columns: min/max for numerics, cardinality for categories, and null rates for missing data. Compare row counts before and after joins to catch unintended duplication.

Quality reviews should explicitly ask who is over- or under-represented and whether collection incentives skew results. The bias glossary entry links to storytelling pitfalls.

Plain-language limitation statements build trust. Point readers to the Data Literacy Basics hub for vocabulary you can reuse in footnotes and slide decks.