Self-created demonstration using synthetic data, not a paid client project.
Problem: a small export contained an exact duplicate, missing fields and inconsistent text. Blind cleanup could lose legitimate records or invent values.
Approach: preserve the source, apply explicit normalization rules, remove only exact duplicates, and keep missing values and conflicting records visible.
Result: the six-record sample retained five records after one exact duplicate was removed. Unknown and missing values were retained and flagged. The workbook separates Source, Cleaned and Change log tabs so the result can be checked against the input.
Deliverables: sample workbook, cleaned preview, source preview and change log. AI-assisted preparation with rule-based checks. This is a small demonstration, not a claim of client savings or production deployment.
Like this project
Posted Oct 26, 2026
Self-created demo: six synthetic records become five after exact deduplication. Source, cleaned data and change log make the result auditable.