Data

Four source families feed every model. Labeling, coverage and matching are method — they live in Methodology · §2.

1 First American — Silver

AuditLIVE Silver AE QA — 28 × 8 counties ↗ · #104 · ARS↔Smarty 91/91 ↗

FA data dictionaries33 sources, field-level Data sourcesproperty · owner · history FA roof codesROOF_COVER + assessor codes

2 BuildZoom — Gold

AuditLIVE Gold-QA — 23 × 13 counties ↗ · #100

Gold variable inventorypermit history, as landed BuildZoom dictionaryall 44 fields ETL walkthroughBronze → Silver → Gold

3External public sources

Auditplanned Externals-QA — required per P1

Demographics catalogACS · Decennial · IRS · HUD · BLS · FCC US demographics cachenationwide parquet, all models Scraping catalogcounty portals, Stage 1

48020REI clients

Auditin progress Olivia step 1.1 — join keys · date grain · deal volume; freezes the label spec

Olivia — what this labelsdeals · leads · appointments · fulfillments How the label is built§2.2 transaction labels · BigQuery, APN-joinable

Cross-source

Feature taxonomyvariables → features + synthetic families