Population estimates and retail spend estimates are the two input layers that drive trade area statistics for each co-location cluster. Both are derived from publicly available data sources and applied at the H3 resolution-7 hexagonal grid level, per the O-D catchment methodology. Together they supply the population and spend axes used by the deterministic ranking system and the V3 catchment ranking methodology.

Population Data

Population estimates are sourced from the WorldPop 2026 100-metre population grid (worldpop.org). WorldPop produces modelled population estimates derived from census microdata, satellite imagery, and dasymetric redistribution. The 100 m resolution places population at the sub-block level, enabling precise trade area delineation.

Processing pipeline

1. Spatial filter: Only grid cells within 150 km of at least one co-location cluster centroid are retained, reducing data volume by approximately 80% while preserving all cells relevant to catchment computation.

2. H3 aggregation: Retained cells are assigned to their containing H3 resolution-7 hexagon and population values are summed. H3 resolution-7 cells have an average area of 5.16 km².

3. Output: census-h3-res7.jsonl — one record per H3 cell with fields {h3, lat, lon, pop, iso}.

Countries covered

United States, Canada, Mexico, Great Britain, Germany, France, Netherlands, Austria, Portugal, Greece, Denmark, Iceland, and Poland — 13 countries as of the current pipeline version.

Spend Data

Spend estimates are synthesised by applying annual per-capita expenditure multipliers by retail category to the population grid. The multipliers are proxies derived from national household expenditure surveys.

Multipliers are expressed in local currency. Cross-country spend comparisons require foreign-exchange normalisation, which is not applied in the current pipeline. Rankings are most meaningful within a single country or within the eurozone.

Retail categories

• Grocery: Supermarkets, hypermarkets, food cooperatives, and food sections of general merchandise retailers.
• Hardware: Home improvement, building materials, and garden centres.
• Wholesale: Members-only warehouse clubs and cash-and-carry retailers.

Processing pipeline

Spend values are computed at the H3 resolution-7 level by multiplying each cell's aggregated population by the per-capita multipliers for its country. Output: cleansed-spend-h3-res7.jsonl — one record per H3 cell with fields {h3, pop, spend_grocery, spend_hardware, spend_wholesale, currency}.

Catchment Aggregation

For each co-location cluster, primary and secondary catchment zones are defined by crow-flies distance rings (see: O-D Catchment Methodology). Population and spend for all H3 cells within each zone are summed to produce the cluster's trade area statistics. These aggregated values are the basis for cross-cluster competitive ranking.

Point-of-interest data

The retail anchor and secondary operator locations that form co-location cluster centroids are sourced from OpenStreetMap contributors under the Open Database Licence (ODbL). Point-of-interest data is distinct from the population and spend layers described above; it provides the geographic seed points from which catchment zones are measured. The full data-attribution statement covering all pipeline layers appears in topic-regional-markets-system. [osm-odbl]

References

• Trade area — Wikipedia, accessed 2026-06-14
• WorldPop Global High Resolution Population Denominators Project — WorldPop, University of Southampton, accessed 2026-06-14
• H3: Uber's Hexagonal Hierarchical Spatial Index — H3 Geo, accessed 2026-06-14

Wikipedia content reproduced under CC BY-SA 4.0.

OpenStreetMap data © OpenStreetMap contributors, licensed under ODbL.

See also

• od-catchment-methodology
• catchment-ranking-methodology-v3
• co-location-methodology

OpenStreetMap data © OpenStreetMap contributors, licensed under ODbL.