PF03: Tested CSV/XLSX Cleaning Pipeline

PF03 is a self-initiated anonymized data-cleaning demonstration built to make every output count auditable.

Cleaning tabular data without explicit rules can silently lose rows or invent business decisions. The input, rejected records, duplicates, and final totals have to reconcile.

I created a Python workflow that validates the sample schema, normalizes agreed values, identifies errors, separates duplicates, and writes three result files plus a summary.

• 21-row synthetic input

• 8 valid, 11 error, and 2 duplicate rows

• Three result files plus a summary

• Seven automated tests

The counts describe this demonstration dataset only. They are not a paid-client outcome and do not predict another file. Production work begins with a sanitized sample and explicit missing-value and duplicate policies.

Self-initiated anonymized sample. The 8/11/2 reconciliation belongs only to the demo dataset; this is not paid or verified client work.