Geographic co-location methodology

@@ -7,10 +7,10 @@ index_group: scoring-and-clustering
type: topic
content_type: topic
quality: complete
short_description: "Five-degree cluster system scoring retail node proximity from Primary, Secondary, and Tertiary target co-occurrence to rank development sites by anchor strength."
short_description: "Predicate-gate tier system scoring retail co-location clusters by anchor composition, catchment rank, and civic presence to rank development sites by anchor strength."
status: stable
bcsc_class: current-fact
last_edited: 2026-07-11
last_edited: 2026-08-24
editor: pointsav-engineering
language_protocol: PROSE-TOPIC
source_refs:
@@ -32,45 +32,27 @@ four-jurisdiction, Spain-inclusive version is also the one consistent with
paper's inclusion of Spain (MITMA data, 58 clusters). This article is the outlier.
**Flagged, not resolved.**

The methodology produces five cluster degrees, each representing a more refined level of anchor co-occurrence. Higher-degree clusters indicate that a geography has attracted multiple categories of institutional commercial investment — a compound validation of market conditions that exceeds the threshold any single anchor produces alone.
The methodology assigns each cluster to one of four tiers by predicate gate, not by an accumulated score. A cluster earns a tier only when every gate required for that tier passes — anchor composition, catchment rank, and civic presence together, not any one factor alone. This gate structure replaced an earlier points-based scale on 2026-05-16; the retired scale, and its numeric labels, no longer describe any current cluster.

## The Five-Degree Cluster System
## The Tier System

**Correction (2026-08-02):** this name collides with a differently-defined "Five-
Degree Framework" in the JOURNAL working paper
[[geometric-site-selection-national-tenancy]] §4.3, which defines its own 1st degree
as "Primary anchor only... maps approximately to T3" — no secondary anchor required
until the 2nd degree, unlike this article's First-Degree definition below (which
requires a secondary anchor at the very first rung). Same name, same "Degree"
vocabulary, same tertiary-proximity threshold, but genuinely different admission
rules at each rung — neither article acknowledges the other. **Flagged, not
resolved.**
Each cluster is anchored by a store from one of four anchor classes — Hypermarket, Lifestyle, Hardware, or Warehouse — and is assigned to a tier by clearing that tier's required gates. The tier names follow the retail property hierarchy used by the International Council of Shopping Centres: **Regional**, **District**, **Local**, and **Fringe**, from highest to lowest. The full tier definitions and gate tables are maintained in the [[gis-cluster-scoring-glossary|cluster scoring glossary]] and [[co-location-tier-nomenclature|tier nomenclature]]; the summary below states what a development analyst needs to read the map.

The five cluster degrees are constructed from two radius thresholds: a tighter threshold applied to Primary and Secondary Target co-occurrence, and a wider threshold applied to Tertiary Target proximity. Each degree builds on the prior.
### Regional and District — the qualifying tiers for site selection

### First- and second-degree clusters
**Regional** clusters combine a Warehouse or Lifestyle anchor with a Hypermarket anchor, rank in the top decile of their country by primary catchment population, and have a regionally classified hospital nearby. This is the highest-conviction tier: the simultaneous presence of multiple independent anchor categories, at a nationally significant catchment scale, confirms that several institutional operators have reached the same conclusion about the geography.

**First-Degree Cluster** — A [[co-location-target-hierarchy|Primary Target]] site (Walmart Supercentre) with either of the two Secondary Targets (Home Depot or Costco Wholesale) located within the Primary-Secondary proximity threshold of each other. This is the baseline Qualified Investment indicator: the co-presence of a volume grocery operator and a major home improvement or membership warehouse operator at that proximity confirms that the geography meets the site thresholds of two institutionally distinct retail operators.
**District** clusters combine a Hypermarket anchor with a Hardware or Warehouse anchor, rank in the top quartile of their country by primary catchment population, and have hospital access within the civic ring. District clusters are qualified candidates: the co-location and catchment conditions are met, but at a sub-regional rather than a nationally significant scale.

**Second-Degree Cluster** — A First-Degree Cluster with either Tertiary Target (a post-secondary institution or a major medical centre) located within the wider Tertiary proximity threshold. The addition of a Tertiary Target introduces a demand source for professional services tenancy that is structurally independent of the retail node — a university or hospital draws professional staff, patient populations, and academic visitors who represent the primary target tenant base for Woodfine Professional Centres.
### Local and Fringe

### Third- and fourth-degree clusters
**Local** clusters contain a Hardware or Warehouse anchor, rank in the top half of their country by primary catchment population, and have any hospital within the civic ring. Local clusters confirm co-location and community-level catchment support without qualifying at District or Regional scale.

**Third-Degree Cluster** — A Primary Target with both Secondary Targets (Home Depot and Costco) within the Primary-Secondary proximity threshold of each other. The requirement that both secondary operators be present at that same tight radius is a materially higher bar than the First-Degree standard. Home Depot and Costco independently validate a geography through their own site selection processes; their simultaneous presence confirms that multiple institutional operators have evaluated the same geography against their independent criteria and reached the same conclusion.
**Fringe** is assigned to any cluster that does not clear the gates for Regional, District, or Local. A Fringe cluster may still show retail co-tenancy; it lacks the catchment reach, composition, or civic support the higher tiers require.

**Fourth-Degree Cluster** — A Third-Degree Cluster with either Tertiary Target within the wider Tertiary proximity threshold. The combination of both secondary operators plus an institutional tertiary demand source represents a geography in which retail, professional, and institutional activities have converged.
## Non-Overlap Radius

### Fifth-degree clusters as highest-conviction sites

**Fifth-Degree Cluster** — A Third-Degree Cluster with both Tertiary Targets within the wider Tertiary proximity threshold. Fifth-Degree Clusters are the highest-conviction sites in the dataset. The simultaneous presence of Walmart, Home Depot, Costco, a post-secondary institution, and a major medical centre within the respective radius thresholds marks a geography that has independently attracted five categories of institutional commercial commitment.

## Radius Calibration

The Primary-Secondary threshold reflects the operational reality of [[power-centres|power centre]] site design: retailers in the same commercial node or corridor are typically clustered close together, and proximity at that scale confirms shared site infrastructure rather than coincidental geographical proximity. The wider Tertiary threshold reflects the broader catchment pattern of institutional employers. A university or major hospital does not anchor a power centre — it anchors a geography, at a scale consistent with the drive-time tolerance of professional services tenants in a [[about-regional-markets-system|Regional Market]] context.

### Periodic recalibration

If Fifth-Degree Clusters come to represent an outsized share of all Primary Target entries in the full dataset, the thresholds are tightened — both the Tertiary and the Primary-Secondary proximity thresholds are reduced. This recalibration mechanism prevents the top cluster tier from becoming too large to be operationally useful as a site selection filter.
Two nearby clusters are compared for overlap using a fixed 3.0 km radius disk centred on each. When the overlap between two clusters' disks exceeds the non-overlap gate's limit, the weaker cluster is held below the tier its composition would otherwise earn — preventing a single strong node from being double-counted as several separate clusters. This single radius convention applies uniformly; it is not adjusted upward or downward by how many clusters currently qualify at a given tier.

## Dataset Output Metrics

@@ -78,7 +60,7 @@ For each Primary Target entry, the geographic co-location output records: the ci

### Ranking and shortlist production

The output is then ranked to balance cluster degree, Primary-Secondary co-occurrence, and the absolute sales per square foot of the Primary Target. The ranking produces the site shortlist from which Woodfine engages real estate professionals in each identified market to assess land availability and development timeline.
The output is then ranked to balance cluster tier, anchor proximity, and the absolute sales per square foot of the Primary Target. The ranking produces the site shortlist from which Woodfine engages real estate professionals in each identified market to assess land availability and development timeline.

## Required Site Count by Jurisdiction