Summary Counts

Bucket counts over every location matching the active filters โ€” not just the current page.

Available beside each browse route:

EndpointCounts
GET /locations/summaryAll location types together.
GET /locations/facilities/summaryFacilities only.
GET /locations/sheds/summarySupply sheds only.
GET /locations/plots/summaryPlots only.

This exists because a page cannot answer "how much of my portfolio is non-compliant?". Cursor-paginated listings report no total at all, and the page-number total is capped. The summary counts the whole filtered set server-side instead.

Same filters, same numbers

Each summary route accepts the same filter[...] grammar and organisation scoping as the list route beside it, and is built from the same status derivation. So a bucket equals the number of rows its own filter returns:

GET /locations/facilities/summary
โ†’ by_compliance.non_compliant = 12

GET /locations/facilities?filter[compliance]=eq:non_compliant
โ†’ 12 matching facilities

page, page_size, sort, include and fields[...] are rejected with a 400 rather than silently ignored โ€” none has a meaning on a bucket count.

Response

{
  "meta": {},
  "links": { "self": "/locations/facilities/summary" },
  "data": {
    "type": "location_summary",
    "attributes": {
      "counted": 128,
      "capped": false,
      "by_compliance": { "non_compliant": 12, "pending": 4, "compliant": 87, "excluded": 2, "not_assessed": 23 },
      "by_location_type": { "facility": 128, "shed": 0, "plot": 0 }
    }
  }
}
FieldMeaning
countedLocations the buckets cover โ€” the whole filtered set unless capped.
cappedThe filtered set exceeded the summary cap (10 000). Both bucket objects are then null.
by_complianceEUDR status counts, matching filter[compliance]. The five buckets sum to counted.
by_location_typeCounts per type. May sum to less than counted (see below).

Two honest caveats

capped withholds the buckets rather than estimating them. Above the cap the matched set is bounded without an ordering, so a split over it would describe an arbitrary prefix โ€” a wrong number, not an approximate one. Narrow the filters and retry.

Each location lands in exactly one bucket, the first that applies: non_compliant (at least one non-compliant plot), pending (no non-compliant plot, at least one awaiting adjudication review), compliant (every non-excluded plot has a compliant result), excluded (the operator excluded every plot from the filing), else not_assessed.

not_assessed is not a verdict, and is broader than "unscreened". It covers locations where some catchment plot has no deforestation result yet (not screened, still running, or failed), locations with no cached roll-up, every plot (a plot holds no roll-up of its own), and locations ingested but not yet linked to a resource pivot, which is where the status resolves its organisation.

by_location_type can sum to less than counted because location_type is nullable: a location whose type never resolved belongs to no bucket.

Prefer the typed routes

When you know the type, use /facilities/summary rather than /summary?filter[location_type]=eq:facility. The typed route applies the type as a route param, which lets the organisation scope start from the typed index instead of enumerating every entity-linked location in the org.

Required scope: locations:read