Data quality is assessed along several dimensions: completeness, uniqueness, validity of formats and ranges, consistency between fields, freshness. Automatic checks, placed in the pipeline, verify these rules and produce a report, to detect a problem before it reaches users.