Sales CSV Cleaning & Quality Audit — Synthetic Demo

CSV Sales Data Cleaning and Quality Audit Self-initiated demonstration · Synthetic data · AI-assisted

I created a reproducible data-cleaning demonstration using 80 entirely synthetic sales records. The source deliberately includes inconsistent text formatting, currency symbols, duplicate orders, missing values and invalid dates.

Using AI-assisted Python, the workflow standardizes supported values, validates required fields and preserves problematic records for review. The run produced 60 accepted orders and a 20-row quarantine containing eight duplicates and twelve invalid records. No missing values were filled with invented facts.

An independent verification script passed 17 checks, including row conservation, unique order IDs, valid date and amount formats, preservation of the original problem records, and agreement with the fixture’s expected results.

The deliverables are the source CSV, cleaned data, quarantine reasons, audit log and documented scripts. The scripts use the Python standard library. The cover displays the results of this actual synthetic run. This is a new portfolio demonstration, not past client work.