Documentation

How Cartélie is delivered

Cartélie runs as an engagement today. We take your operating geography, build a hexagonal score grid from population, roads, points of interest, land use and transport, and hand back the results as a GeoPackage you can open in QGIS — plus a written analysis of where to build next.

Scoring weights are set per engagement. A fuel retailer and a health network care about different things, so the model is tuned to the decision you are actually making rather than applied from a fixed template.

Planned: hosted API

A self-serve platform is in development. Nothing below is available yet — it is published so you can see the intended shape, not as a reference to build against.

plannedScore gridComposite score per H3 cell across a country or city.

plannedRecommendationsRanked candidate sites with score, grade, and reasoning.

plannedCoverage gapsUnderserved areas relative to an existing network.

plannedIsochronesDrive-time catchments around a site or candidate.

plannedDataset uploadIngest your own network or competitor list as CSV / GeoJSON.

Want the shape of the output before it is hosted? See what an engagement produces.

How it works

From raw competitor data to a ranked deployment plan in under five minutes.

Ingest networks

Upload competitor locations (CSV/GeoJSON), your existing footprint, and base map (OSM). Cartélie geocodes and stores everything.

Analyze coverage

Generate isochrone drive-time rings (15, 30, 60 min). Aggregate into H3 hexagon grid. Build competitor density heatmap.

Score sites

Compute deployment score for each location. Factors: demand (population), competition (distance to rivals), accessibility (road quality).

Deliver ranked list

Export PDF dossier, interactive map with all layers, and API access for integrations. All in 2–5 minutes.