Every parcel attribute we publish started life as a column in some county's GIS system — an APN here, a situsfulladdress there, a roll_impvalue somewhere else. Today we're making that lineage visible. Our county documentation pages now include a schema-mapping diagram: a column-by-column picture of how each county's original source data flows into the harmonized parcel schema you download and query.

What the diagram shows

Counties publish parcel data in wildly different shapes. Los Angeles County splits a situs address across nine fields; Norfolk, Virginia keeps deed references in numbered column triplets; nearly every assessor has their own name for "year built." Our pipeline harmonizes all of it into one nationwide schema, and the new diagram documents exactly how, for each county:

• Source → harmonized mappings. Each entity we acquire for the county (parcels, addresses, zoning, reference tables) appears with its original source columns on the left, wired to the harmonized columns they become. Direct column mappings, computed transformations (with the actual expression shown), and geometry each get their own connector style.
• The final assembly. A summary diagram at the top shows how all of the harmonized entities come together into the published parcel table — including which source ultimately feeds each final column.
• Supplemented fields. Some attributes don't come from the county at all. Elevation statistics (elevmin, elevmax, elevavg) are computed from USGS 1-meter DEM data, and flood-zone fields (fldzone, staticbfe, FIRM identifiers) come from FEMA's National Flood Hazard Layer. The diagram labels these with their provenance and their actual fill rate in that county.
• Determined columns. Ground-truth fields like FIPS codes, county and tract names, and dataset version identifiers are stamped deterministically by the pipeline, and are shown as their own source.
• Nothing hidden. Columns the source doesn't provide are listed plainly, as are the source columns we preserve verbatim in each record's extras JSON rather than discarding.

Why we built it

If you're evaluating parcel data for a serious application — title research, risk modeling, site selection — "is this attribute populated?" is only half the question. The other half is "where did it come from?" A yearbuilt that flows straight from the assessor's roll is a different asset than one inferred from somewhere else. The coverage tables on our county pages have always answered the first question with exact fill rates; the schema map now answers the second. Together they give you a complete, honest picture of what you're buying before you buy it.

See it in action

Have a look at the diagrams for Norfolk, Virginia or Los Angeles County, California — scroll to the "Schema mapping" section, and click any diagram to open it full size. You'll find one on every county page where we've acquired and harmonized source data, generated automatically from the same pipeline metadata that drives our processing, so it stays current as counties are re-acquired.

Questions about a specific county's sources, or an attribute you need that you don't see mapped? Reach us at hello@landrecords.us — we're happy to dig in.