Site Selection
docs(site-selection): reconcile co-location-tiering-scoring into the canonical four-tier article -- the T1/T2/T3 points-based compositional scale it documented was retired platform-wide 2026-05-16 and describes no current cluster. Merged its two passages with no equivalent anywhere else (the cluster-count parameter-sensitivity finding, and the full planned per-cluster strength score: design principles, three-driver set, open weighting question, scorecard contents) into catchment-ranking-methodology-v3; did not carry over the retired T1/T2/T3 material or the cluster-formation recap already covered by co-location-cluster-formation and cluster-deduplication-threshold. Source archived with superseded_by, redirect added, 8 inbound references repointed. Also removed an ES-only table row in regional-market-definition.es.md carrying DBSCAN eps/minPts/IoU parameters and a 226-476 sweep range that the EN pair had already had redacted -- restores EN/ES parity
@@ -66,13 +66,12 @@ Site Selection covers the strategy and investment thesis behind "near Power Cent ## Site Scoring and Trade Areas [[co-location-cluster-formation|Cluster Formation]] and [[cluster-deduplication-threshold|the Deduplication Threshold]] describe how a qualifying anchor becomes a candidate cluster and how overlapping candidates are resolved. [[co-location-ranking-system|The Ranking System]], [[co-location-tiering-scoring|Tiering and Scoring]], and [[catchment-ranking-methodology-v3|the V3 Catchment Ranking Methodology]] set out the deterministic scoring algorithm and its current binary-predicate tier engine. [[geographic-co-location-methodology|Geographic Co-Location Methodology]] covers the continental-scale cluster framework behind the method. [[od-catchment-methodology|Distance-Band Methodology]] and [[trade-area-methodology|Trade-Area Methodology]] define the primary and secondary trade areas — and the honest limits of straight-line distance-band measurement — that scoring runs against. [[co-location-cluster-formation|Cluster Formation]] and [[cluster-deduplication-threshold|the Deduplication Threshold]] describe how a qualifying anchor becomes a candidate cluster and how overlapping candidates are resolved. [[co-location-ranking-system|The Ranking System]] and [[catchment-ranking-methodology-v3|the V3 Catchment Ranking Methodology]] set out the deterministic scoring algorithm and its current binary-predicate tier engine. [[geographic-co-location-methodology|Geographic Co-Location Methodology]] covers the continental-scale cluster framework behind the method. [[od-catchment-methodology|Distance-Band Methodology]] and [[trade-area-methodology|Trade-Area Methodology]] define the primary and secondary trade areas — and the honest limits of straight-line distance-band measurement — that scoring runs against. <!-- AUTO-GENERATED MEMBERSHIP: DO NOT EDIT BELOW — regenerate from index_group: site-scoring-and-trade-areas --> - [[co-location-cluster-formation]] — The pipeline step that converts each qualifying anchor store into a candidate co-location cluster; overlapping candidates from a single trade area are resolved by deduplication before ranking inputs are aggregated. - [[cluster-deduplication-threshold]] — The co-location index pipeline deduplicates overlapping clusters that represent the same commercial zone using a fixed, tightly-set proximity threshold, retaining the cluster with the higher secondary operator count. - [[co-location-ranking-system]] — The deterministic mechanics behind cluster ranking on the co-location platform — country-relative percentile ranking, the overlap test between neighbouring clusters, and the tiebreak order applied within a tier. - [[co-location-tiering-scoring]] — Tier scoring for co-location clusters — what the T1–T3 composition tiers measure, what they explicitly do not claim, and why a headline cluster count is a model output rather than a measurement. - [[catchment-ranking-methodology-v3]] — The current predicate-gate system that assigns each co-location cluster to one of four tiers — Regional, District, Local, Fringe — using anchor composition, national catchment rank, civic infrastructure, and spatial independence, introduced May 2026 to replace an earlier composite-score model. - [[geographic-co-location-methodology]] — Predicate-gate tier system scoring retail co-location clusters by anchor composition, catchment rank, and civic presence to rank development sites by anchor strength. - [[od-catchment-methodology]] — Each co-location cluster is assigned two straight-line distance bands — a primary zone within 35 km and a secondary zone from 35 km to 150 km — that determine the population and spend attributed to it.