Filter Mapping Table (Normalized)

This file is the canonical filter map used across adapters.

Rules

• «Alle» / all values on an axis is the unfiltered request, stored as a NULL fact FK — never a filter row. The upstream "select nothing" request (empty BRASK form field, empty BRIS *.ids array, SSB Bygningsbranner dataset) produces a fact with NULL on that axis column. There is no ALL / ALL_BUILDINGS filter standard_code; total slices are read with an omitted get_fire_data axis param (→ <axis>_code IS NULL). The dimension-total category (cross-tab ALL_BUILDINGS / ALL_AGES on categories.standard_code, dim 5/6) is a separate concept — see category-native-to-cat-code.md.
• Standard filter codes are used cross-source wherever comparison is possible:
  • Fire Type: CONFINED / DEVELOPED
  • Building Type: RESIDENTIAL (the only cross-source building_type filter code; "all buildings" is NULL, not a code)
  • Building Age: 0-5 / 6-10 / 11-20 / 21-30 / 31-50 / 51+ / UNKNOWN — BRASK only, no cross-source equivalent
  • Region: no unified standard — geographic systems differ per source (see §4)
• BRASK filters are submitted as ASP.NET multiselect list box fields in a form POST to https://brask.finansnorge.no/. An empty string value means "Alle" (all values selected). Specific selections pass the numeric option id.
• BRIS filter availability depends on the endpoint:
  • /aggregate ("Alle branner og ulykker") — supports Fire Type, Building Type, Region.
  • /restrictedaggregation ("Hvordan bygningsbranner starter") — supports Fire Type and Region only. Building Type is not available.
• SSB has no request-level filter parameters like BRASK/BRIS. Building type is dataset choice (Boligbranner vs Bygningsbranner). Geography on table 12058 is the PxWeb KOKkommuneregion0000 dimension (national EAK or kommune codes) — see §4. SSB has no Fire Type or Building Age.

Read-time (get_fire_data) vs ingest-time requests

Charts: Passing standard_code strings such as RESIDENTIAL or CONFINED into get_fire_data resolves to all matching filters.id rows on that axis for the requested source(s) (postgres-contract-rls-and-keys.md). Passing multiple tokens in the same axis argument (e.g. two native_code values or both a standard_code and a leaf code) resolves to the union of matching ids (OR within that axis). Omitted filter-axis arguments mean only facts where that axis FK is NULL (upstream unfiltered slice). Different axes are combined with AND.

ETL: Adapter POST bodies use upstream-native identifiers (form ids, Matrikkel codes, BRIS mission-type ids, etc.). On ingest, adapters resolve each applied native to filters.id and write that integer onto fact FK columns — never the native string. filters.native_code and standard_code live on the catalog row only.

ETL rule — one native per slice: Each ingest slice resolves to exactly one filters.id per applied axis; the adapter POST body uses that row's native_code (Matrikkel leaf id, BRIS mission-type id, kommune code, etc.), not a rollup standard_code. When the UI selects a parent standard_code (e.g. RESIDENTIAL, COMMERCIAL), resolveFilterAxisTokens expands it to all matching leaf filters.id values before ingest; the BRIS adapter then passes buildingTypeIdsOverride: [native] with one Matrikkel leaf per slice (policeAggregateBuildingParams in lib/data-model/adapters/bris.ts). Fire type and region follow the same rule (toMissionTypeIds, counties.ids). Building age is BRASK-only (filterNativeCode on the form field). BRIS does not expose building age.

Catalog labels: filters.name is a shortened per-leaf UI label (typically the text before the first comma in native_name). filters.standard_name labels standard_code rollups in the UI (e.g. Bolig for RESIDENTIAL). Neither field is an ingest axis token.

BRIS precision: Examples below sometimes show several mission-type ids in one JSON body (e.g. all CONFINED natives). The HeatWaves BRIS adapter ingests one native mission type id per slice and stamps the matching filters.id into fire_type_code — not the BRIS numeric id itself.

BRIS Endpoint Filter Capabilities

---

1) Fire Type (Branntype)

standard_codes: CONFINED | DEVELOPED

Supported by: BRASK, BRIS (all sub-sources). Not supported by SSB.

Definition: CONFINED — fire contained to the object/item of origin (Begrenset branntilløp). DEVELOPED — fire spread beyond it (Utviklet brann).

Brann i skorstein (10175) exists only in the "Alle branner og ulykker" dataset (Police Sub-A). It does not appear in "Hvordan bygningsbranner starter". It is a CONFINED-class chimney fire unique to that dataset.

How Fire Type is applied in the API

BRASK — ASP.NET form POST field. One fetch per fire type:

# CONFINED (Kald)
ctl00$Innhold$lbType = 2

# DEVELOPED (Varm)
ctl00$Innhold$lbType = 1

# All fire types (empty = Alle)
ctl00$Innhold$lbType =

BRIS /aggregate — revisedMissionTypes.ids array in the JSON body. Fetch one slice per Oppdragstype value, then standardize at ingest:

// DEVELOPED only
{ "revisedMissionTypes": { "ids": [10026], "isMissingValue": false } }

// CONFINED only (Police Sub-A — all three confined types)
{ "revisedMissionTypes": { "ids": [10086, 10212, 10175], "isMissingValue": false } }

// CONFINED only (Police Sub-B, Brigade, DSB — no skorstein)
{ "revisedMissionTypes": { "ids": [10086, 10212], "isMissingValue": false } }

BRIS /restrictedaggregation — same revisedMissionTypes.ids field, restricted body:

{
  "revisedMissionTypes": { "ids": [10026], "isMissingValue": false },
  "counties": { "ids": [], "isMissingValue": false },
  "hundredAndTenCentrals": { "ids": [], "isMissingValue": false },
  "includePoliceCauseWithNoMission": false,
  "aggregationQuery": { ... }
}

Ingest iterates all relevant Oppdragstype values, tags each row with the standardized fire type code, and aggregates the slices.

---

2) Building Type (Bygningstype)

standard_codes: RESIDENTIAL ("all buildings" is the unfiltered slice — NULL building_type FK, not a code)

Supported by: SSB (dataset selection only), BRASK (næring), BRIS /aggregate only. Not supported by BRIS /restrictedaggregation.

How Building Type is applied in the API

SSB — no filter parameter. Two separate datasets are queried depending on the desired scope:

• Boligbranner = RESIDENTIAL
• Bygningsbranner = all building fires (including residential) → stored with building_type_code NULL (the unfiltered total), not a filter row

BRASK — ASP.NET form POST field using the næring (industry) dimension. Only the Beboelse value is used for the building type split:

# RESIDENTIAL (Beboelse)
ctl00$Innhold$lbNæring = 0

# All buildings (empty = Alle næringer) → stored as NULL building_type FK
ctl00$Innhold$lbNæring =

Note: næring has 22 industry sector values in total. For the Building Type filter, only 0 (Beboelse) and empty (all) are relevant. The full næring taxonomy is not used as a cross-source filter — it is a BRASK-specific analytical dimension.

BRIS /aggregate — buildingTypes.ids array of Matrikkel leaf-level 3-digit string codes. Only leaf-level IDs work; parent IDs (e.g. "1" or "11") return zero results.

Explore / Utforsk ingest: When the user selects Bolig (or individual leaves), Explore expands to one ingest slice per leaf filters.id, each calling the API with that leaf's native_code only — not the 30-id Bolig list and not standard_code=RESIDENTIAL. The whole-Bolig JSON below is the upstream shape for a single parent rollup request (native_code 0 or non-leaf RESIDENTIAL parent), not what each leaf slice sends.

// RESIDENTIAL (Bolig — 30 leaf IDs from category "1")
{
  "buildingTypes": {
    "ids": [
      "111","112","113","121","122","123","124","131","133","135","136",
      "141","142","143","144","145","146","151","152","159","161","162",
      "163","171","172","181","182","183","193","199"
    ],
    "isMissingValue": false
  }
}

// All buildings (no filter) → stored as NULL building_type FK
{ "buildingTypes": { "ids": [], "isMissingValue": false } }

Records without building type data are excluded when a filter is applied. To also include records with missing building type, set isMissingValue: true.

BRIS Bolig leaf-code map (native_value -> native_name)

Source of truth: scripts/dsb_detailed_site_map.md ("Bolig Leaf IDs").

Ikke bolig (client-only comparison)

Not a filter row or filters.name. «Ikke bolig» is a UI/analytics shorthand: compare the all-buildings total (building_type_code IS NULL) against RESIDENTIAL, or sum non-residential leaf facts (COMMERCIAL, INDUSTRY, OTHER, PUBLIC, UNKNOWN). See building-type-semantics.ts.

BRIS non-RESIDENTIAL Matrikkel leaves (native_value → native_name)

Upstream enumeration reference: NON_RESIDENTIAL_MATRIKKEL_LEAF_IDS in _brisUpstream.constants.ts (derived from https://brannstatistikk.no/api/v1/buildinginfo/types). Each leaf is a separate filters row with its own standard_code bucket — ingest uses one row per slice, never the full list in one POST.

211: Fabrikkbygning
212: Verkstedbygning
214: Bygning for renseanlegg
216: Bygning for vannforsyning, bl.a. pumpestasjon
219: Annen industribygning
221: Kraftstasjon (>15 000 kVA)
223: Transformatorstasjon (>10 000 kVA)
229: Annen energiforsyningsbygning
231: Lagerhall
232: Kjøle- og fryselager
233: Silobygning
239: Annen lagerbygning
241: Hus for dyr, fôrlager, strølager, frukt- og grønnsakslager, landbrukssilo, høy-/korntørke
243: Veksthus
244: Driftsbygning for fiske og fangst, inkl. oppdrettsanlegg
245: Naust/redskapshus for fiske
248: Annen fiskeri- og fangstbygning
249: Annen landbruksbygning
311: Kontor- og administrasjonsbygning, rådhus
312: Bankbygning, posthus
313: Mediebygning
319: Annen kontorbygning
321: Kjøpesenter, varehus
322: Butikkbygning
323: Bensinstasjon
329: Annen forretningsbygning
330: Messe- og kongressbygning
411: Ekspedisjonsbygning, flyterminal, kontrolltårn
412: Jernbane- og T-banestasjon
415: Godsterminal
416: Postterminal
419: Annen ekspedisjons- og terminalbygning
429: Telekommunikasjonsbygning
431: Parkeringshus
439: Annen garasje- hangarbygning
441: Trafikktilsynsbygning
449: Annen veg- og trafikktilsynsbygning
511: Hotellbygning
512: Motellbygning
519: Annen hotellbygning
521: Hospits, pensjonat
522: Vandrerhjem, feriehjem/-koloni, turisthytte
523: Appartement
524: Campinghytte/utleiehytte
529: Annen bygning for overnatting
531: Restaurantbygning, kafébygning
532: Sentralkjøkken, kantinebygning
533: Gatekjøkken, kioskbygning
539: Annen restaurantbygning
611: Lekepark
612: Barnehage
613: Barneskole
614: Ungdomsskole
615: Kombinert barne- og ungdomsskole
616: Videregående skole
619: Annen skolebygning
621: Universitets- og høgskolebygning med integrerte funksjoner, auditorium, lesesal o.a.
623: Laboratoriebygning
629: Annen universitets-, høgskole- og forskningsbygning
641: Museum, kunstgalleri
642: Bibliotek, mediatek
643: Zoologisk og botanisk hage
649: Annen museums- og bibliotekbygning
651: Idrettshall
652: Ishall
653: Svømmehall
654: Tribune og idrettsgarderobe
655: Helsestudio
659: Annen idrettsbygning
661: Kinobygning, teaterbygning, opera/konserthus
662: Samfunnshus, grendehus
663: Diskotek
669: Annet kulturhus
671: Kirke, kapell
672: Bedehus, menighetshus
673: Krematorium, gravkapell, bårehus
674: Synagoge, moské
675: Kloster
679: Annen bygning for religiøse aktiviteter
719: Sykehus
721: Sykehjem
722: Bo- og behandlingssenter, aldershjem
723: Rehabiliteringsinstitusjon, kurbad
729: Annet sykehjem
731: Klinikk, legekontor/-senter/-vakt
732: Helse- og sosialsenter, helsestasjon
739: Annen primærhelsebygning
819: Fengselsbygning
821: Politistasjon
822: Brannstasjon, ambulansestasjon
823: Fyrstasjon, losstasjon
824: Stasjon for radarovervåkning av fly- og/eller skipstrafikk
825: Tilfluktsrom/bunker
829: Annen beredskapsbygning
830: Monument
840: Offentlig toalett

---

3) Building Age (Bygningsalder) — BRASK only

standard_codes: 0-5 | 6-10 | 11-20 | 21-30 | 31-50 | 51+ | UNKNOWN

Supported by: BRASK only. Not supported by SSB or BRIS.

BRASK exposes 13 fine-grained age buckets. HeatWaves normalizes these into 6 coarser buckets for storage.

How Building Age is applied in the API

Building Age is a BRASK-specific dimension that can be used in two ways:

1. As a filter — restrict the query to a single age bucket, then retrieve Kilde or Årsak data.
2. As a row or column dimension — retrieve a breakdown of fire counts by age bucket.

BRASK — ASP.NET form POST field:

# Single age bucket (6-10 år)
ctl00$Innhold$lbAldersgruppe = 3

# Multiple buckets (select both 6-10 and 11-15)
ctl00$Innhold$lbAldersgruppe = 3
ctl00$Innhold$lbAldersgruppe = 4

# All age groups (empty = Alle)
ctl00$Innhold$lbAldersgruppe =

Normalization note: two BRASK native values (Ny and 1-5) both map to the 0-5 normalized bucket. When querying by normalized bucket, the adapter must send both native ids to cover the full range.

BRIS — not supported on any endpoint.

SSB — not supported on this axis.

---

4) Region

No unified standard code — geographic systems are incompatible between sources.

Supported by: BRASK (cresta-sone, pre-2020 counties), BRIS /aggregate and /restrictedaggregation (counties.ids and municipalities.ids), SSB table 12058 (KOKkommuneregion0000 — national EAK or per-municipality codes; see ssb-pxwebapi-tables.md and pipeline §4.3).

SSB — table 12058 (KOKkommuneregion0000)

• Alle regioner / national: PxWeb EAK → stored as region_code IS NULL (not a filters.id).
• Filtered kommune: { "filter": "item", "values": ["<4-digit code>", …] } → region_code = resolved filters.id (same spine as BRIS municipality ids where codes align).

BRASK — Cresta-sone (pre-2020 county boundaries)

BRASK uses Cresta-sone ids on form field ctl00$Innhold$lbFylke (empty = Alle). Native values and Cresta → canonical 2024 county mapping: regions-and-cresta.md §1.

Note: Cresta id 13 is absent in upstream numbering (see regions doc).

BRIS — Counties and Municipalities

BRIS uses current Norwegian administrative geographic units. County IDs and municipality IDs come from the national geographic registry (SSR/Kartverket). These do not correspond to BRASK's Cresta-sone ids.

Examples:

• Oslo = county id 3, municipality id 301
• Viken = county id 30 (post-2020 merged county, contains old Akershus, Buskerud, Østfold)

The full county and municipality ID lists are fetched from GET /api/v1/buildinginfo/nace (or via the Pinia store where-data-options) at runtime and are not hardcoded in HeatWaves.

How Region is applied in the API

BRASK — ASP.NET form POST field. Empty = all regions. Single or multiple values:

# Oslo only
ctl00$Innhold$lbFylke = 3

# Oslo and Akershus
ctl00$Innhold$lbFylke = 3
ctl00$Innhold$lbFylke = 2

# All regions (empty = Alle)
ctl00$Innhold$lbFylke =

BRIS /aggregate and /restrictedaggregation — counties.ids (coarse) or municipalities.ids (fine) array in the JSON body. Both endpoints accept this field. Empty array = all regions:

// Single county (Oslo = 3)
{ "counties": { "ids": [3], "isMissingValue": false } }

// Single municipality (Oslo = 301)
{ "municipalities": { "ids": [301], "isMissingValue": false } }

// All regions
{ "counties": { "ids": [], "isMissingValue": false } }

---

Cross-Source Comparability

Maintenance

When changing BRASK form fields or BRIS county/municipality filters, update this file and regions-and-cresta.md. Adapter code: lib/data-model/adapters/_registry.ts, _braskWebForms.constants.ts, bris.ts.