Spec portfolio project using generated fictional data — not client work.
I built an Excel data-entry and reconciliation control from four fictional receipts. The workflow preserves source identifiers and exact money fields, converts the sources into seven traceable line items, keeps one duplicate visible while excluding it once, flags ambiguous text for review, and reconciles the accepted total to $151.52 with a $0.00 discrepancy.
The artifact set includes the fictional source images, a source contact sheet, the controlled Excel workbook, an exception log, rendered review pages, a manifest, and visual-review evidence.
Verification covered quantities, unit prices, tax, line totals, receipt subtotals, duplicate handling, ambiguity flags, and the grand total. Ten artifacts passed 16 independent checks.
AI-assisted tools supported implementation and QA; the final specimen was reviewed and tested. No client, revenue, operational, or business outcome is claimed.
Synthetic receipt-to-Excel control with preserved source IDs, visible duplicate and ambiguity handling, exact reconciliation, and documented QA. Not client work.