Steps
-
Select approved exports and confirm owner, population, cutoff and encoding. Do not add credentials or connect this starter to production systems.
-
Run profile_csv.py on the included synthetic file first. Review the HTML dashboard and JSON results.
-
Run on an approved local CSV with a documented key. Record schema and quality observations in the workbook.
-
Validate business meaning and materiality with the source and consumer owners. Use the Actions sheet to assign decisions.
Synthetic example
The synthetic loan file includes one duplicate loan ID, a missing status and an invalid amount. The profiler identifies technical observations; it does not decide the correct business value.
Scope
The profiler reads only the file path supplied, runs locally, and makes no network calls. It reports row duplication, key duplication, missingness, inferred types and numeric parse issues. It is not a mainframe parser, an enterprise crawler or a security scan. Treat inferred types as provisional.