Normalizing seller tapes without the spreadsheet marathon
Every seller sends a different tape. Here's how automated schema mapping turns days of column wrangling into a reviewable, repeatable process.
- Written by
- Capital Markets DeskLoanDNA
- Published
- Reading time
- 6 minutes
- Topic
- Capital Markets

Loan tapes are the lingua franca of secondary markets, and yet no two speak the same dialect. Column names drift, codes change meaning between sellers, and dates come in every format imaginable.
Mapping once, reusing forever
LoanDNA learns a seller's tape structure the first time it sees it and proposes a mapping into a unified schema. Analysts confirm or adjust, and the mapping is saved for every future tape from that counterparty.
- Suggested field mappings with confidence and sample values
- Code translation tables for occupancy, purpose and property type
- Validation rules that catch out-of-range or impossible values
From normalized to scrubbed
Once normalized, every loan on the tape can be compared to the documents in its file. The gaps between what the tape says and what the documents prove are exactly where pricing risk hides.
