#71data analysis
Audit Data Quality Before Reporting
Finds missing, duplicate, inconsistent, and invalid records.
Base prompt
Inspect this dataset for [intended report]. Profile columns, types, uniqueness, nulls, ranges, dates, and relationships. Identify duplicates, impossible values, inconsistent labels, and suspicious changes over time. Show issue counts and example rows without exposing private identifiers. Propose validation rules and safe corrections; never silently discard records. Deliver a concise issue log ranked by impact on the report.Follow-up steps
- 1
Refine and verify
Tests the output against the original goal and improves its practical usefulness.
step 1Re-run the profile after proposed corrections and compare each issue count and the report totals before and after.
Comments0
No comments yet.
