From Raw Parcel Data to a Ranked Research Queue
A parcel universe counts in the millions, yet an analyst can only meaningfully review a few dozen candidates a day. This walkthrough traces the five moves that turn raw records into a ranked, source-attributed research queue: discovery, verification, identity resolution, reduction, and priority ranking — with priority meaning readiness to be studied, never a prediction that a deal will close.
How does raw parcel data become a ranked research queue?
The pipeline runs in stages. Discovery pulls parcels and ownership from authoritative records. Verification attaches a source and a date to each material field and marks what is unknown. Identity resolution collapses records that describe the same physical asset into one candidate. Reduction filters out anything that cannot yet be studied. Finally, ranking orders the survivors by research priority — how ready each one is for analyst attention given the evidence on hand — so the shortest, most defensible worklist rises to the top.
The problem is not a shortage of parcels; it is a shortage of analyst hours. A county's records may hold hundreds of thousands of properties, but only a fraction match a mandate, and fewer still carry enough current evidence to justify a phone call. Moving from that raw universe to a worklist an analyst can act on is a sequence of narrowing steps, each of which either adds provenance or removes ambiguity.
Discovery and verification
Discovery begins with authoritative records rather than listings or scraped aggregates. County appraisal districts and assessors maintain parcel-level ownership and physical characteristics as public record, and those files anchor the candidate universe. Verification then walks each material field — parcel identifier, owner of record, land and improvement descriptors — and stamps it with the record it came from and the date it was collected. A field the source does not provide is stored as unknown, never guessed. This discipline is what lets an analyst later trust a value without re-pulling it by hand.
- Each field carries a citation to the record that produced it and the date of collection.
- Missing values are represented as unknown rather than filled with a plausible default.
- Conflicts between two records for the same field are flagged, not silently averaged.
- Freshness is tracked so an analyst can see how old the underlying data is.
Identity resolution and reduction
A single apartment complex can appear as several parcels, under variant owner spellings, or across two datasets with mismatched formatting. Identity resolution decides when separate records describe one physical asset and merges them into a single candidate, keeping the provenance of every contributing field. Reduction then removes candidates that cannot yet be studied: those whose evidence is too thin, too stale, or too internally contradictory to defend. What remains is a set of candidates that each stand on inspectable facts.
| Stage | Input | Output |
|---|---|---|
| Discovery | Authoritative parcel and ownership records | Raw candidate universe |
| Verification | Raw candidates | Fields with source, date, and unknown markers |
| Identity resolution | Verified records | One candidate per physical asset |
| Reduction | Resolved candidates | Candidates with defensible evidence |
| Ranking | Defensible candidates | Ordered research queue |
How ranking is scored
Ranking orders candidates by research priority: a composite of how much material evidence is present, how current it is, and how cleanly the identity resolved. A candidate with fresh, unconflicted, fully attributed fields sorts above one with gaps or stale sources. The score answers a single question — which candidate is most ready to be examined next — and deliberately says nothing about whether an eventual deal is likely to close.
- —A high rank means research-ready, not that the owner will sell or that the asset is a good buy.
- —The queue reflects only the sources onboarded so far; a market not yet ingested will not appear.
- —Stale or conflicting public records limit how far reduction can narrow a candidate, and those limits are shown rather than hidden.
Frequently asked questions
Why rank on research priority instead of a close probability?
A close probability would imply a conclusion the public record cannot support. Research priority stays inside what the evidence can defend: it measures readiness to study, leaving the investment judgment to the analyst.
What happens to candidates that get filtered out?
They are not discarded permanently. When a fresher record arrives or an ambiguity resolves, a previously reduced candidate can re-enter the queue with updated provenance.
See a ranked queue built from your target counties.