Product

An evidence warehouse for property and infrastructure diligence.

Siboga stores the raw public record, parses it, resolves the parcel and the owner, keeps the history, and distributes pinned snapshots. It is built for teams that must show their work.

Operator lead lists for real-estate investors are served by a separately licensed product. Siboga is the warehouse and the institutional channel.

Printed wrapper of the first Siboga-Expeditie monograph, 1901

What you can buy

What is on offer, each with a gate before sale.

Property event feed

Funds, lenders, and acquisition teams working a defined geography.

A historical snapshot plus scheduled deltas for foreclosure, auction, tax delinquency, permit, and assessment events. Source link and observation time on every event.

Gate: Sold only where source-by-source redistribution has been reviewed and match rates are measured.

Evidence dossier

Infrastructure investors, counsel, corporate development, and investigative teams.

A bounded answer to one asset or ownership question, with a source ledger, claim grades, unresolved paths, and reproducible evidence.

Gate: Every material conclusion gets a second reviewer. No unsupported negative claims.

Source and coverage audit

Data teams, insurers, and proptech companies deciding what to build.

Jurisdiction by jurisdiction: what exists, how it is accessed, posted terms, fields, cadence, failure modes, and commercial-use status.

Gate: Public access and redistribution rights are reported as separate questions.

How a record is built

Raw, normalized, derived. Kept separate.

Raw

The response body as the source gave it, stored write once with request shape, response headers, fetch time, and SHA-256. The normalized layer must be reconstructible from raw alone.

Normalized

Typed, deduplicated rows with a reference back to the raw object. Each event is matched to a parcel by parcel ID first, coordinates second, and address last, and the match method and confidence are stored on the row.

Derived

Joins, flags, and cross-signal chains computed from normalized data. Always recomputable, never hand edited. A wrong derived value is fixed in the derivation, not in the row.

Fields

What a row carries.

FieldMeaning
signalWhich pipeline produced the row: foreclosure, pending sale, tax delinquent, permit, or assessment.
property_addressNormalized street address. Used for matching, never as a key.
townMunicipality of the parcel.
ownerOwner of record from the assessor. No contact fields.
event_typeWarehouse event type, for example foreclosure_filed or tax_delinquent.
event_dateWhen the source acted: filing date, due date, sale date, application date.
observed_atWhen Siboga fetched the record.
docketCourt docket for foreclosure and auction rows.
balanceUnpaid balance read from the bill detail page.
oldest_unpaid_yearEarliest unpaid tax year on the bill.
sale_datePosted auction date.
permit_idPortal permit identifier.
use_codeAssessor land-use code.
absenteeWhether the mailing address differs from the property address.
source_urlThe page the row was read from.

The full dictionary, including permit and assessment fields, ships with every snapshot.

Refusals

What Siboga will not sell.

  • No national coverage claims. Coverage is listed town by town, with the query that produced each count.
  • No scores before backtesting. A distress or motivation score needs historical outcomes and a documented baseline first.
  • No contact data. Applicant emails and phone numbers in permit sources are excluded from every export, API, and file.
  • No negative findings without a denominator. "No record found" names the sources searched, the date range, and the sources not searched.