Chaos
180 rows with duplicate keys, whitespace, case drift, mixed countries, currencies and dates, invalid quantities and genuinely missing lead times.
Lilambd Sample Project / data reconstruction · spreadsheet QA
Data Cleanup & Audit
A synthetic supplier-price export becomes a deduplicated, import-ready dataset with raw evidence, transformation rules, QA actions and unresolved values preserved.
state transformation
Operation
Raw preservation → deterministic normalization → composite-key deduplication → validation → exception quarantine → reconciliation.
Usable state
A buyer can import 150 standardized rows, inspect 30 removals and review every value that could not be safely completed.
uncompressed residual
What the model does not erase.
Ambiguous identity, unknown source authority and missing values remain visible rather than guessed.
observable proof
180 raw rows → 150 delivered rows
30 duplicates removed with audit trail
five editable delivery sheets
return to proof library