Skip to content

Place-History Projection

A tool — instantiates Event-Log-Centered Modeling

Assembles the full history of a place by gathering every event bound to it into one time-ordered account, resolving the many names a single place goes by.

Ask "what has happened here?" and a current-state record answers with today's owner, today's zoning, today's occupant — and erases everything that produced them. Place-History Projection reconstructs the answer the record threw away: it gathers every event in the log that is bound to a given place and assembles them into one time-ordered history of that place. Its distinguishing problem — the one that makes it more than a location filter — is that a single place wears many names across a long log (an address renumbered, a parcel subdivided and remerged, a building renamed), so the projection must resolve place identity before it can gather a coherent history. It answers the spatial question the timeline and the entity-trajectory views leave open: not what one actor did, nor what happened when, but the whole biography of a location.

Example

A city's land-records office needs the full history of one parcel before a contaminated-site sale can close. The event log holds a century of happenings — deeds, subdivisions, permits, zoning variances, an underground-tank installation, an environmental remediation — but they reference the parcel under four different lot numbers and two street addresses, because the parcel was split in 1962 and partly remerged in 1998. Place-History Projection first resolves those references to one place-identity, stitching the pre-split and post-merge fragments into a single subject. Then it gathers every bound event and orders them, producing a chain of title and use: who held it, what was built, what was buried, what was cleaned.

The output is a single readable dossier — and where the record is silent (a fourteen-year gap with no recorded events), the projection shows the gap rather than papering over it, because "no event" is not the same as "nothing happened." That distinction is exactly what a contaminated-site buyer needs: the visible gap is where the environmental due-diligence must go looking.

How it works

  • Resolve place identity first. Reconcile the variant references — addresses, lot numbers, informal names — that denote one place, and split apart references that look alike but denote different places, so the gathered history belongs to exactly one subject.
  • Gather every bound event. Pull all events whose place-and-context binding points at the resolved place, across every type — ownership, construction, use, incident.
  • Order into a biography. Sequence the gathered events in time to form the place's trajectory, carrying each event's context (who, what, under what conditions).
  • Show the silences. Leave recorded gaps visible instead of interpolating, so absence of evidence is not read as evidence of absence.

Tuning parameters

  • Place-resolution strictness — how confidently two references must match before they are treated as the same place. Loose matching stitches a fuller history but risks fusing two neighbouring parcels; strict matching keeps places clean but fragments one place into several.
  • Place granularity — the level a "place" is defined at (a room, a building, a parcel, a district). Fine grains give precise local histories; coarse grains give context but blur exactly-where.
  • Context breadth — how much surrounding context travels with each event (adjacent parcels, weather, jurisdiction). Broad context enriches the history; narrow context keeps it focused and fast.
  • Boundary-change handling — how splits, merges, and renumberings are threaded. You can follow the land through boundary changes or follow the legal parcel; the two give different, both-valid histories.

When it helps, and when it misleads

It is the right tool wherever the place is the subject of the question — title and due diligence, incident investigation at a site, provenance of an artifact by location, or asset histories where a physical thing accumulates events over decades. Because it is derived, a better place-resolution rule yields a better history on re-run, without touching the events themselves.

Its central risk is silent place-resolution error. Fuse two parcels and the history contains events that never happened there; fragment one parcel and half its past vanishes — and both failures are invisible in the tidy output unless resolution confidence is surfaced.[1] The subtler trap is reading a gap as calm: an empty stretch usually means the log wasn't capturing events for that place, not that the place sat inert, so an incomplete log quietly understates a location's history. The disciplines are to carry place-resolution confidence into the projection, to keep record gaps explicit, and to state the log's coverage for the place so a reader knows how much of the true history the dossier can contain.

How it implements the components

Place-History Projection realizes the spatial slice of the archetype — binding events to places and resolving which place is which — and leaves timing machinery and relational linking to siblings:

  • place_and_context_binding — it reads and gathers on exactly this binding; the projection is nothing but the set of events bound to a resolved place, in order.
  • entity_identity_resolution_rule — applied to places: reconciling the many references (addresses, lot numbers, names) that denote one location, and separating look-alikes, before any history can be assembled.

It does not build participation or causal edges among events — that is Event Knowledge Graph — and it does not define the generic projection contract or rebuild path, which is Event-Sourced Projection.

References

[1] In property law a chain of title is the ordered succession of recorded conveyances establishing who has held a parcel; a break or ambiguity in the chain is a defect that must be cured. It is the classic real-world place-history, and its discipline — every link evidenced, every gap flagged — is exactly what this projection must reproduce for any kind of place. ↩