Data licensing
Every source in the Kortex graph carries an explicit commercial status; stated here, on the source register, and in the provenance envelope of API responses. Where rights are not yet confirmed, the status says so. We would rather show an honest “under review” than an implied “cleared”.
Production-safe today
201 of 233 exposed sources are production-safe today; their licence question is settled (cleared, derived-only, or licence held) and you can build commercial products on query results drawn from them, subject only to the stated attribution conditions. The remaining 32 sources are under review and are not production-safe until their status changes here.
Policy: raw records from under-review or excluded sources are not served on anonymous public surfaces (the Global Viewer demo layers, homepage demo queries, and static downloads). Enforced 2026-07-03: live vessel (AIS) and aircraft (ADS-B) feeds are key-gated pending the provider-rights audit, and the nuclear-reactor, cable-landing and volcano map layers were withdrawn; each action is recorded in the source notes below. Keyed API responses identify every source they touched in the provenance envelope, so a query that used an under-review source says so explicitly.
Execution modes: the rights position as a control, not just a disclosure
Every envelope-bearing endpoint accepts ?rights_mode=:
409 rights_mode_blocked. Evaluation requires an explicit opt-in per request or per agreement.provenance envelope. The public demo runs in this mode; pilots choose their mode in the agreement.Example blocked response:
{"error": {
"code": "rights_mode_blocked", "status": 409,
"message": "This calculation draws on sources that are not commercially cleared; rights_mode=commercial_safe fails closed rather than serving them.",
"details": {"blocking_sources": ["telegeography"],
"hint": "Re-run with rights_mode=evaluation to receive the answer with full source disclosure, or check /licensing for clearance status."}}}
Status definitions
The source licence permits commercial use and redistribution. Any conditions (attribution, share-alike) are stated in the notes and honoured in Kortex’s response provenance.
Redistribution rights not yet confirmed for commercial use. The data remains in the register; the note states what is exposed today and what is being resolved.
By policy, no raw records from this source are served. Only Kortex-derived indicators computed from it appear on any surface, and the source licence permits that derived use.
Kortex holds a commercial licence from the publisher covering this use.
Ingested for internal evaluation only. Not exposed through the API, the graph, or any commercial surface.
The matrix
| Source | Licence | Status | Notes |
|---|---|---|---|
| ADB Projects | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| adsb.lol community ADS-B | CC0 inbound; outbound unstated (open-data posture) — derived-only serving | derived only | Feeders contribute under CC0 and no use restriction is stated (terms captured 2026-07-19); absent an explicit outbound licence, serving stays derived-only with attribution "data from adsb.lol". Raw per-aircraft serving held. |
| AESO Alberta pool price and ETS public data | AESO legal notice - non-commercial/personal/educational grant only | excluded | DETERMINED 2026-08-29 (T143b; evidence raw/tso_terms_20260829/aeso/): the only grant is "non-commercial, personal or educational purposes"; any other use needs AESO's written permission. EXCLUDED fail-closed pending that permission (ST-45 founder letter). Tension recorded - pool price is statutory public data under Alberta's EUA, but the operative site terms do not reflect that; the letter should cite it. |
| AidData | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| AIS vessel positions (aisstream.io) | none published (free service, key-gated by operator) — derived-only serving | derived only | Terms reviewed 2026-07-19 (captures in raw/aisstream/) - no terms of service exist on the site, documentation states no use restriction, and the operator's GitHub explicitly welcomes company use of the free API. Absent an explicit outbound data licence, anonymous surfaces serve derived aggregates only (counts + density, no vessel identifiers) and raw per-vessel layers stay key-gated. |
| ANEEL RALIE - Relatorio de Acompanhamento da Expansao da Oferta de Geracao de Energia Eletrica | ODbL 1.0 - attribution "ANEEL - Agencia Nacional de Energia Eletrica" | cleared | CLEARED 2026-08-12 (lead 726, batch_determinations_20260812_evening.txt; evidence pack raw/br_aneel_ralie/ckan_package_show_20260812.json). Licence read from the publisher's OWN CKAN package_show field - license_id "odc-odbl", license_title "Open Data Commons Open Database License (ODbL)", organization "Agencia Nacional de Energia Eletrica" - the portal IS the publisher, the strongest evidentiary shape available. The triage text-match that first flagged this lead ("licenca ambiental") was a FALSE POSITIVE, an environmental-permit data VALUE inside the rows rather than a licence statement; the ruling rests on the CKAN field, not that match. ODbL precedent in this estate - osm_planet, openinframap, icij_offshore, pypsa_eur_base - share-alike carried in this registry entry, so this attribution string must never be merged with a plain CC-BY source's. DELIBERATELY SKIPPED (recorded per house rule) - the three *-historico.zip/.parquet vintage siblings (retrievable later for lineage, not needed for the fleet column) and the three PDF data dictionaries (reference only). REACHABLE THROUGH THE US RELAY, consistent with the rest of the BRA/ARG corridor. FORMAT - semicolon-delimited, double-quoted, CRLF, genuine UTF-8 (verified on multi-byte accented sequences, not latin-1/cp1252 mojibake), no BOM. EVERY column stored as TEXT including the Brazilian comma-decimal Mda* fields and the many Dat* date fields - casting a subset would mean silently asserting which columns are "real" numbers, the kind of claim the ar_energia loader explicitly avoided for potencia_instalada_mv. NO GEOMETRY in any of the three resources; CodCEG (ANEEL's generation-enterprise code) is the nearest spatial join key for a future traversal. |
| Atlas of Economic Complexity | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Australia national electricity infrastructure (Geoscience Australia) | CC BY 4.0 - attribution (c) Commonwealth of Australia (Geoscience Australia) 2026 | cleared | CLEARED 2026-08-24. Closes the gpkg_375 Australia deferral - the AURIN gatekeeper route stayed dead (WFS times out from this network, re-confirmed 08-24) but the deferral's premise expired, GA serves the same data from its own ArcGIS server, found via the ArcGIS Online catalog rather than data.gov.au. Rights chain explicit PER LAYER, not inferred - substations cleared in batch_determinations_20260812 lead 480, transmission lines and major power stations carry their own CC-BY-4.0 CKAN records (au-govt-ga-national-electricity-transmission-lines-2016-na, au-govt-ga-national-major-power-stations-2016-na), and the service copyrightText '(c) Commonwealth of Australia (Geoscience Australia) 2026' shows active maintenance beyond the 2017 snapshot the lead pointed at. TRANSPORT - the server answers 403 to plain clients and 200 to a browser User-Agent, same class as AusTender ATM. Count gate per layer refuses ArcGIS silent truncation (the gpkg_375 Nigeria lesson). |
| Australian Energy Market Operator (NEMWEB) | AEMO Copyright Permissions; general permission for any purpose with attribution | cleared | CLEARED 2026-08-13 with attribution (determination and captures in raw/aemo_nemweb/; corporate site is Cloudflare-gated so the policy text was evidenced via the Internet Archive 2026-07-29 capture, then confirmed verbatim first-hand by the founder same day, ST-31 retired). AEMO confirms general permission for anyone to use AEMO Material for any purpose with accurate and appropriate attribution; NEMWEB PUBLIC files are AEMO-created, publicly made available, and marked PUBLIC in each file's own C record. Attribution carried per row - "(c) AEMO, DispatchIS_Reports, nemweb.com.au". Non-binding constraint rows and non-DispatchIS report families are recorded skips, refetchable. |
| Austrian cadastre (BEV Kataster DKM) | CC BY 4.0 (attribution wording conventional; BEV publishes no prescribed credit string) | cleared | Cleared 2026-08-02 from the BEV ISO19139 record, verbatim 'Fuer dieses Produkt gilt die Standardlizenz CC-BY-4.0' and 'Der oeffentliche Zugang zu diesem Produkt unterliegt keinen Einschraenkungen'. Area NOT carried - FLAECHE lives in the Grundstuecksverzeichnis at a different grain (15.4M rows over 10.2M parcels) and is not inferred. Geometry is EPSG:31287 projected, so no axis hazard; the Rechts-/Hochwert attributes in sibling tables are label-placement only and are not read. |
| Awesome Electrical Grid Mapping (Open Energy Transition) | CC0 1.0 Universal | cleared | The INDEX is CC0 1.0 (LICENSE captured verbatim at raw/grid_source_leads/), so reusing the index itself carries no obligation and no attribution duty. That says NOTHING about the licences of the datasets it points at - each of those needs its own rights read against the publisher own terms before anything is ingested, and grid_source_leads.verified stays false until one is done. |
| Bank for International Settlements | BIS terms (reproduction with acknowledgment) | cleared | BIS statistics may be reproduced with source acknowledgment (verified 2026-07-11). Attribution carried on /sources and /licensing. |
| Belgian cadastre (SPF Finances / AAPD plan cadastral) | SPF Finances Licence plan cadastral opendata (CC BY 4.0 structure, non-sublicensable) | cleared | Cleared 2026-07-31 on the federal monthly product, which covers all three regions under ONE licence (determination raw/be_cadastre/). Commercial use, redistribution and the sui generis database right are all granted; downstream restrictions are forbidden; the licence is NON-SUBLICENSABLE. Attribution has no mandated string but a mandated element list (author, copyright notice, licence reference, warranty-limitation reference, URI) plus a duty to state modification. TWO CLAUSES BIND DOWNSTREAM USE - Art 7(c) bars use in a decision based solely on automated processing producing legal or significant effects on a person, and Art 7(d) requires GDPR compliance; neither restricts ingest or analysis. RULED 2026-08-15 (Stu delegated the call as a common-sense one; ST-08 retired) - Kortex outputs are corporate-focused screening and decision-SUPPORT for organisations about land and grid; no surface makes or presents an automated determination about a natural person, no natural-person data from this source is served (parcel fabric only, and natural-person tenure is barred estate-wide by the T111 corporate-first rule), and verdict language must keep it that way; Art 7(d) is covered by the standing GDPR posture (/data-protection, T07). Any future surface that WOULD touch natural persons re-opens this ruling before shipping. THE WALLOON REGIONAL COPY IS CONTRACTUALLY NON-REDISTRIBUTABLE and is deliberately not used, even though its REST layer returns geometry unauthenticated - being fetchable is not being licensed. The public domain is largely NOT cadastred (~5% of territory), so roads, rivers and rail have no polygon. |
| Berlin cadastre (Geoportal Berlin, ALKIS Flurstuecke) | Datenlizenz Deutschland Zero 2.0 (dl-de/zero-2-0); no attribution condition | cleared | Licence read from the service's own ows:Fees, verbatim 'Fuer die Nutzung der Daten ist die Datenlizenz Deutschland - Zero - Version 2.0 anzuwenden.' ZERO so attribution is not a condition. FORMAT TRAP: the WFS GeoJSON output is NOT RFC 7946 lon/lat - it declares urn:ogc:def:crs:EPSG::25833 in a crs member and returns UTM33N eastings/northings, so a loader trusting the container format would read 402438 as a longitude and PostGIS would accept it silently; the loader refuses to run if the service ever declares a different EPSG. FIELD TRAP: bezeich reads 'AX_Flurstueck' on every row - it is the AAA feature-type name, constant, and carrying it as a designation would fill 403,484 rows with the same string. No INSPIRE CadastralParcels endpoint and no bulk file exist for Berlin, so paging the WFS is the intended route. |
| BIG / Satu Peta Indonesia - electricity infrastructure layers | no express licence - service root asserts copyrightText "Sekretariat Kebijakan Satupeta" with no grant | excluded | INTERNAL_ONLY 2026-08-12 (determination raw/id_big/id_big_rights_determination_20260812.txt). THE LEAD THAT WOULD HAVE BEEN THROWN AWAY - all three layers carried NO licence tag in the source index, so under the old habit they would have been dismissed from across the internet without a byte fetched. Downloaded and triaged locally under Stu's 08-12 rule instead. WHAT THE LOCAL READ FOUND - the MapServer root asserts only copyrightText 'Sekretariat Kebijakan Satupeta' and the per-layer metadata carries an EMPTY description and no copyright field, so the publisher asserts attribution and grants nothing expressly. THE ANSWER CAME BACK BARRED ANYWAY, and that is the rule working rather than failing - it changed WHERE we look for a licence, never the fail-closed default, and internal_only is now a finding against the publisher's own metadata rather than an assumption about a missing tag. Indonesia's Satu Data policy (Perpres 39/2019) is a publication mandate not a third-party licence and is NOT relied on. WORTH HOLDING EVEN BARRED - Indonesia is 48.8 GW and had NO grid data of any kind before this; layer 9 gives 1,109 substations EVERY ONE carrying kapgi (capacity) and teggi (voltage), layer 11 gives 2,961 plants with daya in MW and primary energy type. UNITS ARE THE PUBLISHER'S AND NOT OURS - teggi arrives as text ('150 kV') and stays text; kapgi and daya are bare numbers with no unit stated anywhere in the service metadata, so they keep the publisher's column names with a note recording the unit is UNSTATED. Naming a column capacity_mva would assert a unit they never gave, the exact fault behind three unit incidents here. Z DIMENSION on the line layer is forced off rather than the column widened - an export artifact, not surveyed elevation, and stated because dropping a dimension is a decision. |
| Bready economy assessments | CC BY 4.0 | cleared | World Bank B-READY data, CC BY 4.0. Attribution carried on /sources, /licensing and in response provenance. |
| CAIDA IODA Internet Outages | None published | excluded | IODA publishes no data licence (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| CAISO OASIS market data | CAISO OASIS terms | cleared | Public OASIS API, no key; CAISO terms review in progress; fails closed in commercial_safe; on-row licence carried. |
| Cambridge CBECI | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Caparéseau — capacités d'accueil S3REnR (FR) | Caparéseau published capacities (indicative; site terms) | cleared | Cleared 2026-07-28 after terms review; the Mentions Legales data carve-out states portal data is public and reuse permitted with source citation. Mandatory attribution 'Source - capareseau.fr (RTE)'; figures served verbatim, indicative character disclosed. Determination + captures raw/capareseau/terms_review_20260728.txt. Applies to all operators delivered through the portal. |
| CEH/NERC EIDC GB electricity substations (33kV+), 2018 | OGL v3 - attribution "Contains data supplied by UK Centre for Ecology & Hydrology" | cleared | CLEARED 2026-08-13 (batch_determinations_20260813.txt group 3 CEH_EIDC, lead 4594 - the sole grid-relevant standout of that group's 27 members; the remaining 26 are ecological/hydrology/ agronomy science datasets with no grid-siting use and were NOT registered). OGL v3 read from the record's own "Licence and reuse information" section. THE GIVEN URL IS THE SUPPORTING- DOCUMENTS BUNDLE ONLY (readme + ro-crate-metadata + a data-sources docx, no substation rows); the real data file is declared inside that bundle's own ro-crate-metadata.json as data/gb_substations_data_281118.csv, resolved via the DataDownload distribution URL's HTML interstitial to the complete package zip - followed programmatically, not assumed. UTF-8 WITH BOM (utf-8-sig), same trap ar_energia hit. Columns are the publisher's own headers, mechanically snake-cased only (space to underscore) - CONSTRAINT_STATUS included - never translated. |
| China photovoltaic plant vector datasets (ChinaPV + CPVPD-2024) | CC BY 4.0 - attribution per dataset | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt, leads 168 and 170). Both Zenodo records state license id cc-by-4.0 and access_right open. CORRECTS T91's "CHN has ZERO open" verdict, reached without fetching - BUT THE SUBSTANTIVE CONCLUSION SURVIVES, both are photovoltaic PLANT datasets (generation assets), not transmission topology, so China's grid network remains unheld. ChinaPV ships 2015 (1,645) and 2020 (10,985) vintages loaded here (a third file, a labelled test-sample subset, was NOT loaded - it is a methodology artefact, not a third vintage); CPVPD-2024 ships as 34 per-province shapefiles (11,636 features) inside one zip. Loaded via fiona, geometry stored as-is (Polygon/MultiPolygon mixed, column typed generic Geometry) with source vintage/province carried per row. |
| CISA ICS Advisories | US Public Domain | cleared | Public domain; no restrictions. |
| Climate TRACE | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Companies House bulk data | Crown copyright — CH free data products (public register) | cleared | Companies House free data products are public register information made available for re-use free of charge. Attribution carried on /sources, /licensing and in response provenance. |
| Copernicus CAMS | Copernicus open licence | cleared | Copernicus licence; free use including commercial, with attribution. |
| Correlates of War | COW terms (non-commercial, no redistribution) | excluded | COW terms prohibit commercial use and redistribution (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| Croatian cadastre (DGU, Digitalni katastarski plan) | Otvorena dozvola (attribution DGU with date of last change) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/croatia_terms_20260806.txt) - the Otvorena dozvola grants every use expressly including commercial, reproduction, distribution and integration into products; conditions are source attribution WITH the date of last change and marking modifications. TWO WFS GENERATIONS TRAP - the old ZIS DKP WFS is 'Ograniceno koristenje' (registered users); only the new anonymous INSPIRE/HVD endpoints and the ATOM channel may be used - being fetchable is not being licensed. A sample KO zip was measured at build (fields CESTICA_ID, BROJ_CESTICE, POVRSINA_GRAFICKA, MATICNI_BROJ_KO); gpkg_344. |
| Czech cadastre (CUZK, INSPIRE CadastralParcels) | CC BY 4.0 (attribution CUZK, 2026) | cleared | Cleared 2026-07-31; CUZK provides the data 'bezuplatne na zaklade licence Creative Commons CC-BY 4.0' and names the INSPIRE GML files in scope. ATTRIBUTION IS PRESCRIBED - 'uvedeni zdroje ve formatu - CUZK, [rok]', so exactly 'CUZK, 2026', and modified works must state the modification. TWO AXIS TRAPS - the EPSG:4258 variant is latitude-first, and 5513 vs 5514 Krovak would yield transposed and mirrored geometry; the 5514 variant is used and was verified against a known point. Boundary accuracy varies (surveyed DKM vs vectorised KMD); cp:estimatedAccuracy is published per boundary but not yet carried. Fabric never ownership - Czech ownership data is the separately governed VFK format. |
| Danish cadastre (Matriklen jordstykker via DAWA) | CC BY 4.0 frie grunddata (credit Geodatastyrelsen) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/denmark_terms_20260806.txt) - Datafordeler register-specific terms for Matriklen state CC BY 4.0 for frie grunddata with verbatim credit duty 'Du skal kreditere Geodatastyrelsen pa et passende sted'. Terms note that the publisher may change conditions at any time - re-read on registration if the Datafordeler route is ever taken. DAWA serves WGS84; declared CRS asserted per response in gpkg_341. |
| Data-centre layer (PeeringDB + Kortex hyperscaler register) | per row (PeeringDB terms / Kortex-curated) | cleared | Constituents cleared/curated; per-row source discloses origin. |
| Derived grid topology (OpenInfraMap + OpenStreetMap) | ODbL (both constituents; share-alike) | cleared | Both constituents cleared ODbL with share-alike; attribution OpenInfraMap + OpenStreetMap contributors. |
| DOE EAGLE-I county outage archive | CC BY 4.0 with required depositor citation - verified, and NOT a US Government Work public-domain release | cleared | DETERMINED 2026-08-31 (T142). THE LOAD-CRITICAL SEMANTIC, gated not assumed - 1,972,650 rows (6.47 percent) carry an EMPTY customers_out, which means NO READING from the utility map, not "nobody is out". Loading blank as zero would fabricate all-clear readings across 6.5 percent of the record including during outage events, so blanks land NULL and the loader REFUSES to swap if the NULL count ever collapses to zero. Two disclosures carried on the data - the publisher's own coverage and DQI quality instruments stop at 2022 while the data runs to 2025, so any share-of-customers claim for 2023+ has an unmeasured denominator; and the LIVE EAGLE-I system is out of scope by its terms (DOE/Federal staff, contractors and state emergency coordinators only, "for official use only"), so only this published archive is used. GEOGRAPHY SEAM NAMED - 8 of 3,076 FIPS do not join tiger_counties, and they are precisely Connecticut's eight legacy counties (09001-09015) which the Census replaced with nine planning regions for statistical purposes in 2022. EAGLE-I still reports legacy county FIPS. Those rows load and are queryable by FIPS, but carry no county geometry until a legacy-to-planning-region crosswalk is added - recorded rather than silently dropped. |
| DOE OE-417 electric disturbance events | US Public Domain | cleared | US government work; public domain. |
| Dutch cadastre (Kadaster BRK Kadastrale Kaart) | CC BY 4.0 (attribution Kadastrale kaart (c) Kadaster) | cleared | Cleared 2026-07-30 on the data.overheid.nl catalogue record, which declares CC-BY 4.0 at dataset level and on every distribution (capture raw/nl_kadaster/dataoverheid_licence_20260729.json, determination terms_review_20260730.txt). Attribution; Kadastrale kaart (c) Kadaster, CC BY 4.0. |
| E-Redes Open Data (PT) | E-Redes Open Data (portal terms) | under review | Portal licence terms to be verified for commercial redistribution — under review; fails closed in commercial_safe mode. |
| EBRD Projects | EBRD website terms (non-commercial) | excluded | EBRD site terms are personal/non-commercial (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| EC GHSL (Global Human Settlement Layer) | Free use with attribution (CC BY 4.0) | cleared | European Commission GHSL data policy - free use including commercial, with attribution. |
| ECMWF ERA5 | Copernicus open licence | cleared | Copernicus licence; free use including commercial, with attribution. |
| EDF-SEI grid hosting capacity, French island systems | Licence Ouverte / Open Licence 2.0 (Etalab) - attribution EDF-SEI | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt, lead 334). Licence Ouverte 2.0 read from the data.gouv.fr record's own license field (lov2). BLOCKED ON TWO PRIOR ATTEMPTS (T102, T103) - the old domain opendata-corse-outremer.edf.fr answers every advertised Opendatasoft API path, including its own exports endpoints, with a 302 redirect or a ~147KB HTML shell. RESOLVED 2026-08-13 - the whole EDF open-data estate has migrated to a new unified portal, opendata.edf.fr, on a different platform (Koumoul data-fair, not Opendatasoft); the old domain now hard-redirects there for every request, and the new platform answers plain HTTP JSON with no bot gate. Found by searching the new platform's own dataset catalogue for the French title the old record used and reading off the combined dataset id (Capacites d'accueil du reseau - Corse & Outre-Mer). SCOPE CORRECTION CARRIED FROM T102, STILL TRUE - this is EDF-SEI (Systemes Energetiques Insulaires), the NON-INTERCONNECTED ISLAND SYSTEMS, NOT mainland France; mainland hosting capacity (Enedis/RTE) remains unheld. NO GEOMETRY - poste is a substation name, not a coordinate, so no place gate applies. |
| EIA US Energy Atlas (substations) | Public domain (US government work; HIFLD lineage) | cleared | US federal public domain; no gating, attribution courteous. |
| Electricity North West open data (Opendatasoft) | CC BY 4.0 | cleared | CC BY 4.0 — attribution carried on-row and on /sources. |
| Elexon Insights (GB Balancing Mechanism) | Elexon BMRS open data licence (attribution "Contains BMRS data © Elexon Limited copyright and database right 2026") | cleared | Licence reviewed and cleared 2026-07-15. BMRS open data licence grants worldwide, royalty-free, perpetual, non-exclusive use incl. commercial exploitation in own products. Attribution statement + licence link carried on /sources, /licensing and in response provenance; no implied official status or Elexon endorsement. Licence page archived at review date (capture 2026-07-03). |
| Ember | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Endangered Languages Project | ELP terms (non-commercial) | excluded | ELP terms are non-commercial (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| ENTSO-E Transparency Platform | ENTSO-E Transparency licence | cleared | ENTSO-E Transparency Platform legal notice permits reuse with source acknowledgment. Attribution carried on /sources, /licensing and in response provenance. |
| Environment Agency open data (CAMS + Flood Map for Planning) | OGL v3 | cleared | Open Government Licence v3 via the Defra Data Services Platform; commercial use permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| EPA CAMPD (CEMS) | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). |
| ERCOT (Texas ISO) Public API | ERCOT Public API / Terms of Use | cleared | Terms reviewed and cleared 2026-07-18. ERCOT website Terms of Use s.5 - publicly available raw data may be used, reproduced and redistributed in compilations, charts and analyses. The API Terms of Use add access limits and a three-downloads-per-report-per-year rule (respected by design: each report is fetched once and re-served from local storage) and impose no redistribution restriction. No implied ERCOT endorsement; ERCOT marks not used. Terms pages archived at review date (capture 2026-07-18). |
| ERCOT station alias crosswalk (ERCOT + EIA + OpenInfraMap) | per row (constituent licences) | cleared | All three constituents individually cleared; per-row columns carry the applicable one. |
| ERT (Episodes of Regime Transformation) | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| ESB Networks availability-capacity heatmap (IE) | ESB Networks publication (no explicit open licence) | under review | Published quarterly for connection customers with a use disclaimer; no explicit open licence found on the publication page — under review until terms are confirmed with ESB Networks. Fails closed in commercial_safe mode. |
| Estonian cadastre (Maa- ja Ruumiamet, katastriuksused) | Avaandmed; attribution Maa- ja Ruumiamet with as-of date | cleared | Cleared 2026-08-07 on the MTP data-service terms (determination raw/intl_cadastre_recon/estonia_terms_20260806.txt), clause 2.5 verbatim permits use 'nii arilisel kui mitte-arilisel eesmargil, sh teha andmed uldsusele kattesaadavaks' - commercial and non-commercial use including making the data publicly available. Attribution obligation is a dated citation, publisher's own example 'Katastriuksuste piirid Maa- ja Ruumiamet 06.08.2026'. No registration anywhere on the route. CRS L-EST97 EPSG:3301, artefact-gated in gpkg_337. |
| EU Consolidated Financial Sanctions List (FSF) | CC BY 4.0 / Commission Decision 2011/833/EU (reuse for commercial purposes, without charge, without application) | cleared | DETERMINED 2026-08-29 (evidence raw/eu_sanctions_fsf/) - the data.europa.eu record publishes the download URLs including the FIXED public token (token-2017), so no credential is held; licence field COM_REUSE resolves to Decision 2011/833/EU quoted verbatim in the evidence (reuse commercial, free, no application) with the personal-data caveat recorded - person rows are NOT loaded (enterprise screen only), which keeps the GDPR surface at nil. Closes the T142 screen-becomes-CLEARED half-met item - the primary screen now covers OFAC + OFSI + EU. The loader leg existed since 07-11 (load_eu_fsf in gpkg_237) waiting only on the token discovery. |
| EU ETS / Union Registry compliance data (EEA) | EEA standard re-use (free with attribution) | cleared | EEA data policy - free re-use with attribution. |
| European Central Bank SDW | ECB reuse policy (free, attribution) | cleared | ECB permits free reuse including commercial with attribution (verified 2026-07-11). Attribution carried on /sources and /licensing. |
| Eurostat | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution (verified 2026-07-11). |
| Eurostat regional accounts (NUTS) | CC BY 4.0 | cleared | Eurostat reuse policy / CC BY 4.0; attribution carried. |
| Eurostat/GISCO Census 2021 population grid, 1 km, V3 | CC BY 4.0; attribution Source Eurostat (Census 2021 grid), (c) European Union | cleared | Cleared 2026-08-16 (determination raw/eurostat_census_grid/eurostat_census_grid_rights_determination_20260816.txt). CC BY 4.0, commercial reuse without restriction with attribution. GISCO administrative boundaries carry extra restrictions; this statistical grid does not. |
| EV adoption (OWID / IEA Global EV Outlook) | CC BY 4.0 (OWID processing; IEA underlying; cite both) | cleared | CC BY 4.0 with attribution to OWID and IEA. |
| Exelon PHI hosting-capacity maps (Pepco, DPL, ACE) | Utility hosting-capacity map terms | under review | Public ArcGIS FeatureServer, no key; utility map terms review in progress; fails closed in commercial_safe; on-row licence carried. |
| FEMA National Flood Hazard Layer (US flood zones) | US Government Work (usa.gov/government-works) - but see the route caveat in the note | under review | MEASURED BUT DELIBERATELY NOT INGESTED 2026-08-29 (T142). The DATA is public domain (US Government Work, data.gov record captured) - but every FEMA-origin host (hazards.fema.gov, msc.fema.gov, fema.gov) refuses ALL estate egresses including the US relay and TLS impersonation (cloud-ASN block, evidence captured). The one practical route is an Esri Living Atlas mirror (annually-refreshed reduced set, March 2026 snapshot) whose hosting terms are the Esri Master License Agreement and whose item is flagged Retiring - PD data through an encumbered pipe. Fail-closed until either (a) a rights read clears extraction from the Esri mirror, or (b) a non-cloud US egress reaches FEMA's own services. Route decision recorded on the register. |
| FEMA National Risk Index | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). Ingested via the official FEMA ArcGIS feature service (fema.gov hosts block this VPS). |
| FERC EQR (Electric Quarterly Reports, via PUDL) | FERC public record; PUDL processing CC BY 4.0 | cleared | US federal public record; PUDL processed tables CC BY 4.0 with attribution. |
| FERC Form 1 | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). |
| Finnish cadastral index map (MML kiinteistorekisterikartta) | CC BY 4.0 (attribution MML + dataset + delivery date) | cleared | Cleared 2026-08-07 on licence (determination raw/intl_cadastre_recon/finland_terms_20260806.txt) - MML open data is CC BY 4.0 with a THREE-PART attribution duty quoted verbatim in the capture: licensor name, dataset name AND delivery date. Access is registration-gated but openly licensed, the same posture Denmark's Datafordeler holds. No account is created by agents (house rule). |
| Fragile States Index | CC BY-NC-SA 4.0 | under review | CC BY-NC-SA 4.0 (verified 2026-07-02). FSI scores are exposed on Jurisdiction records; resolving via publisher permission or derived-only posture. |
| FRED (St. Louis Fed) | FRED API terms | under review | FRED terms permit use; redistribution of certain third-party series is restricted; series-level audit pending. |
| Freedom House | CC BY-NC 4.0 | under review | CC BY-NC; one country-year score currently exposed on Jurisdiction records; resolving via publisher permission or a derived-only posture. |
| French cadastre (Etalab PCI vecteur) | Licence Ouverte 2.0 (attribution Cadastre - Etalab / DGFiP) | cleared | Cleared 2026-07-29; the data.gouv.fr dataset record declares licence fr-lo and the Licence Ouverte 2.0 text permits reuse for any purpose including commercial, worldwide and royalty-free, on attribution alone (captures raw/fr_cadastre/, determination terms_review_20260729.txt). Attribution; Cadastre - Etalab / DGFiP, Licence Ouverte 2.0. |
| GB DNO network lines (SSEN + NPG + UKPN portals) | per row (three DNO open-data licences) | cleared | Each row carries its own DNO licence (SSEN/NPG/UKPN open data licences, all cleared individually). |
| GEBCO | Public Domain | cleared | Public domain; no restrictions. |
| geoBoundaries | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| GeoNames | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| German Land cadastres via INSPIRE CadastralParcels | per Land; Rheinland-Pfalz is dl-de/by-2-0 with a PRESCRIBED credit string | cleared | RLP licence read AT THE SERVICE, not a catalogue - ows:Fees states 'geldleistungsfrei; Datenlizenz Deutschland - Namensnennung - Version 2.0', AccessConstraints NONE. Attribution is PRESCRIBED and propagates to downstream users, reproduce verbatim including the acquisition year and the modification marker. AXIS TRAP CENTRAL TO THIS LOADER - INSPIRE publishes EPSG:4258 latitude before longitude; a verified RLP posList opens 49.016 8.042 which is Rheinland-Pfalz but read unswapped plots in Somalia. The swap is asserted against the Land bounding box before any write. |
| German Land cadastres, AdV ALKIS shapefile bulk products | Per Land, carried on the row - Bremen CC BY 4.0 (attribution "Landesamt GeoInformation Bremen (LGV Bremen), CC BY 4.0") | cleared | REGISTERED 2026-08-29 by the integration audit (T150). The loader gpkg_335 declared @source_id de_adv_shp and recorded freshness under that id since 2026-08-02, but the registry half of the register-every-source rule was never done - so freshness_watch, which joins strictly on registry ids, could not see it (a dead write). Each Land's licence is read from the publisher's own product page and carried per row; Bremen is CC BY 4.0. No tier is asserted here beyond what the loader already records. |
| German Land cadastres, INSPIRE CadastralParcels WFS | Per Land, read verbatim from each service's own ows:Fees - Hessen dl-de/zero-2-0 (no attribution condition); Brandenburg dl-de/by-2-0 (PRESCRIBED "GeoBasis-DE/LGB"); Sachsen dl-de/by-2-0 (PRESCRIBED "Geodaten Sachsen"); Schleswig-Holstein CC BY 4.0 (PRESCRIBED "(c) GeoBasis-DE/LVermGeo SH/CC BY 4.0") | cleared | REGISTERED 2026-08-29 by the integration audit (T150), same defect as de_adv_shp - loader recorded freshness under an unregistered id since 2026-08-02, invisible to freshness_watch. Licences are NOT inferred: each was read from the service's own ows:Fees at harvest and is carried verbatim per row, including the three PRESCRIBED credit strings (GeoBasis-DE/LGB, Geodaten Sachsen, GeoBasis-DE/LVermGeo SH/CC BY 4.0) which are reproduce-verbatim obligations, not paraphrasable. Sachsen-Anhalt was deliberately excluded by the loader. |
| GLEIF | CC0 | cleared | CC0; no restrictions. |
| Global Dam Watch | CC BY 4.0 | cleared | Global Dam Watch database is distributed CC BY 4.0 (publisher site + Scientific Data descriptor, verified 2026-07-02). Attribution carried on /sources, /licensing and in response provenance. |
| Global Energy Monitor | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| GloHydroRes global hydropower plant and reservoir dataset | CC BY 4.0 | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt, lead 19). Zenodo record states license id cc-by-4.0, access_right open. Global hydropower reservoir/plant linkage - generation, not grid, relevant to the generation fleet rather than the network. Loaded from the publisher's own GloHydroRes_vs1.csv. |
| GRanD (Global Reservoir and Dam DB) | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Greek cadastre (Ελληνικό Κτηματολόγιο / Hellenic Cadastre, operating-cadastre parcels) | CC BY 4.0 (HVD Reg 2023/138 statutory floor; no per-dataset badge - basis stated) | cleared | Cleared 2026-08-13 (lead 9151/GRC, determination raw/lead_determinations/batch_determinations_20260813.txt, evidence raw/intl_cadastre_recon/greece_recon_20260813.txt) - the dataset page names EU Implementing Regulation 2023/138 (High-Value Datasets, HVD category Γεωχωρικά/Geospatial) and Directive 2019/1024 as applicable legislation, access rights "Δημόσια - ανοικτά δεδομένα" (public open data); portal terms 3.4 name CC0 and CC BY 4.0 as the portal's licence types. NO per-dataset licence badge exists - recorded honestly, the ruling rests on the named statutory instrument (the Poland art. 40a precedent: statutory free access, not a badge). Attribution: Ελληνικό Κτηματολόγιο (Hellenic Cadastre) via data.gov.gr. The cadastre's own domains (ktimatologio.gr, data.ktimatologio.gr) are Akamai-gated 403 to every UA; the bulk route survives via the separate data.gov.gr host. LOAD CONDITION: gpkg_397 inspects every regional zip for embedded LICENSE/README/terms-like members before insert and aborts on any contradiction; findings captured to raw/eu_parcels/gr/embedded_terms/. |
| Henley Passport Index | Henley & Partners terms (proprietary) | under review | Henley & Partners terms require a licence for commercial use (verified 2026-07-02). Openness scores are exposed on Jurisdiction records; removal or replacement under evaluation. |
| Hessen statutory flood zones (HQ100) | CC BY 4.0 | cleared | Ruling 2026-08-13 (lead_determination_pack_20260813.md group G2, lead 5488). Licence read from the shapefile's own bundled HLNUG_INSPIRE_Nutzungsbedingungen.txt: "Die kostenfreie Nutzung der Geo- und Metadaten sowie der Geodatendienste ist fuer private und kommerzielle Zwecke erlaubt. Lizenz: Creative Commons BY 4.0." No prescribed attribution string beyond the source credit already carried on the row. |
| HIFLD (Homeland Infrastructure Foundation-Level Data) | US Public Domain | cleared | Public domain; no restrictions. |
| HM Land Registry CCOD / OCOD | HMLR Free Use licence (dataset-specific, NOT OGL) | under review | Free Use licence permits use within the business but prohibits bulk republication/sub-licensing; raw title stores are POLICY-HELD (never servable via /v2/tables); only derived per-counterparty diligence evidence serves. Commercial terms review pending; recorded 2026-07-14. |
| HMLR INSPIRE Index Polygons | OGL (HMLR + Ordnance Survey 100026316 attribution required) | cleared | OGL with two required attribution statements, carried in check limitations and on this register; conditions at use-land-property-data.service.gov.uk/datasets/inspire/#conditions. Attribution; This information is subject to Crown copyright and database rights 2026 and is reproduced with the permission of HM Land Registry; polygons Crown copyright and database rights 2026 Ordnance Survey 100026316. Recorded 2026-07-16. |
| HMLR Price Paid Data (England & Wales) | OGL v3 (address fields carry Royal Mail PAF third-party rights) | cleared | Cleared 2026-08-15 with a FIELD-LEVEL carve-out (determination raw/hmlr_ppd/determination_20260815.md). OGL v3 grant names commercial use expressly. The publisher states the address fields derive from AddressBase Premium incorporating Royal Mail PAF - any commercial use of the ADDRESS DATA beyond residential property price information services needs Royal Mail. Posture - serve price/date/type/tenure/town-grain facts and derived per-parcel context (basis address_match); NEVER republish paon/saon/street/locality+postcode combinations on a commercial surface. DERIVED VEHICLE 2026-09-03 - land_value_postcode (ingest/land_value_postcode_build.sql) aggregates sales to POSTCODE grain at UPRN-centroid for the /v2/spatial/land-value layer; address fields deliberately dropped at build, so the carve-out is satisfied by construction. |
| Hofstede Insights | Hofstede terms (research free; commercial requires permission) | under review | Free for research; commercial use requires permission from Geert Hofstede BV (verified 2026-07-02). Dimension scores are exposed on Jurisdiction records; permission being resolved. |
| HydroSHEDS / HydroRIVERS | HydroSHEDS v1 licence (free incl. commercial, attribution) | cleared | HydroSHEDS v1 licence permits commercial use with attribution (verified 2026-07-11). |
| Hyperscaler datacentre registry | Kortex-curated (public announcements) | cleared | Kortex-curated register compiled from operators' public announcements; first-party compilation. |
| IAEA NFCIS | IAEA terms | under review | IAEA Nucleus terms permit use in commercial products with acknowledgment; confirming no database-specific restriction applies to this registry. |
| IAEA PRIS | IAEA terms (educational/research use) | under review | IAEA Nucleus terms permit use in commercial products with acknowledgment; confirming no database-specific restriction applies to this registry. |
| IAEA Research Reactor DB | IAEA terms | under review | IAEA Nucleus terms permit use in commercial products with acknowledgment; confirming no database-specific restriction applies to this registry. |
| ICE daily hub prices via EIA wholesale markets page | proprietary (ICE); EIA republishes "with permission" | excluded | ICE-originated proprietary series; EIA's permission does not extend to us. Held, not served. |
| ICIJ Offshore Leaks | ODbL | cleared | ODbL; commercial use permitted; attribution required and share-alike applies to derivative databases. Query results are served as produced works with ICIJ Offshore Leaks Database attribution. |
| IESO Ontario public reports and market data | IESO Terms of Use - limited non-exclusive reproduction licence with a PRESCRIBED verbatim copyright notice | cleared | DETERMINED 2026-08-29 (T143b; evidence raw/tso_terms_20260829/ieso/): express licence to use and reproduce conditioned on the exact notice "Copyright (c) 2004-2022 Independent Electricity System Operator, all rights reserved. This information is subject to the Terms of Use set out in the IESO's website (www.ieso.ca)" - the string is PRESCRIBED, not paraphrasable (the Northern Powergrid / BTN100 trap class), and must ride with served reproductions. No non-commercial restriction. |
| IMF Data | IMF open data | cleared | IMF open data terms; used with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| INSEE Filosofi 2021 donnees carroyees, carreaux 200m | Licence Ouverte v2.0 (Etalab); attribution Insee, Filosofi 2021 | cleared | Cleared 2026-08-16 (determination raw/insee_filosofi/insee_filosofi_rights_determination_20260816.txt). LO 2.0 permits commercial reuse with attribution. Guadeloupe/Guyane/Mayotte absent at the publisher. |
| IRENA | IRENA open data | cleared | Published as IRENA open data; used with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Irish connection registers (EirGrid + ESB Networks) | EirGrid / ESB Networks publications (no explicit open licence) | under review | Operator publications with use disclaimers; no explicit open licence found — under review until terms are confirmed. Fails closed in commercial_safe mode. |
| ISO New England | ISO-NE website Terms and Conditions | derived only | ISO-NE website Terms and Conditions (captured 2026-07-23) are a reserve-all disclaimer with no express data-use grant and a hedged duplication caveat; the Information Policy governs protected areas only, not these public CSVs. Derived-only posture; scope of derived context under review. |
| Italian cadastre (Agenzia delle Entrate, INSPIRE) | CC BY 4.0 (attribution Agenzia delle Entrate) | cleared | Cleared 2026-07-30; the publisher's bulk-download page states verbatim 'Il servizio di download massivo della cartografia catastale e disponibile con licenza CC-BY 4.0' (determination raw/it_catasto/). Open because cadastral parcels are an EU high-value dataset under the Open Data Directive. CARVE-OUT - the autonomous provinces of Trento and Bolzano run their own cadastres under delegation and are NOT in this product, so Italian coverage discloses the gap. AXIS TRAP - EPSG:6706 publishes latitude before longitude; coordinates are swapped on ingest. Parcel fabric, never ownership; the visure ownership register is a separate fee-bearing service and is not used. |
| ITU ICT statistics | ITU terms | under review | ITU data terms; redistribution scope unconfirmed. |
| JAO Core Publication Tool (flow-based market coupling) | JAO publication data (public API, no key; site terms) | under review | Public publication tool, no registration; redistribution terms not formally cleared — moot while use stays validation-only (derived statistics about our own model, no JAO data served). |
| Japan MOJ registry-map parcels (G-Spatial GeoJSON conversions) | MOJ terms (commercial use explicitly permitted) + prescribed G-Spatial conversion attribution + no-quality-guarantee caveat - all three travel together, verbatim in parcel_source_registry row 71 | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/japan_conversions_terms_20260807.txt; MOJ licence PDF archived raw/jp_moj_cadastre/license_20260802.pdf) - MOJ 利用規約 operative sentence grants free use including 商用利用 (commercial use); the conversions add no new restriction, only (1) the prescribed attribution 「登記所備付データ○○市」(法務省)(当該ページのURL)をもとにG空間情報センターにて変換処理して作成 and (2) the no-quality-guarantee caveat 「データの内容・品質は保証できません」 which is a DISCLOSURE OBLIGATION wherever this source is described, with conversion errors acknowledged by the publisher. CKAN per-dataset licence tags are operator noise (cc-by 1,941 / cc-zero 14 / gfdl 9 / cc-by-sa 1) - dataset notes defer to the MOJ terms. Count-vs-estimate ratio logged per municipality by gpkg_350 = load-phase validation of the recon-measured ~4 percent conversion-fidelity question. Raw MOJ XML route (login-gated, one vintage ahead) reserved as a future corroborating lineage. |
| JEPX Japan spot market prices | JEPX disclaimer - use permitted with source clearly indicated (shutten wo meiji) | cleared | DETERMINED 2026-08-29 (T143b; evidence raw/tso_terms_20260829/jepx/ original Japanese + gloss): copyright JEPX, use permitted "with the source clearly indicated"; no non-commercial or redistribution restriction stated. CLEARED with attribution. REGISTER-RULE NOTE: this source (and ieso/aeso/nordpool) had been feeding tso_prices UNREGISTERED since July - found and fixed by this audit. |
| JRC Global Surface Water v1.4 | EC free reuse with attribution | cleared | EC/JRC open data - free use with attribution (cite Pekel et al. 2016). |
| Kortex country crosswalk | derived | cleared | First-party Kortex derivation. |
| Kortex derived analytics | derived | cleared | First-party Kortex derivation. |
| Kortex internal staging | internal | cleared | First-party Kortex derivation. |
| Kortex raster tile cache | derived | cleared | First-party Kortex derivation. |
| Kortex TSO aggregator | derived | under review | First-party aggregation over 53 TSO feeds; per-feed redistribution terms are tracked individually (ENTSO-E cleared; remaining feeds under audit). |
| Kortex Voronoi tessellations | derived | cleared | First-party Kortex derivation. |
| Kummu et al. gridded GDP & HDI | CC0 | cleared | CC0; no restrictions. |
| Latvian cadastre (VZD, Kadastra informacijas sistemas atvertie telpiskie dati) | CC BY 4.0 (attribution Valsts zemes dienests) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/latvia_terms_20260806.txt) - CKAN license_id CC-BY-4.0 and the VZD page states verbatim that the open data carry no commercial or non-commercial use restrictions. LICENCE SPLIT TRAP recorded at recon - some geolatvija HVD Atom feeds are CC0 while this national SHP set is CC BY; pin per-feed if the Atom channel is ever used. CRS LKS-92 TM EPSG:3059 read from the artefact .prj in gpkg_338. |
| LBNL Berkeley Interconnection Queue | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Lithuanian cadastre (Registru centras, kadastriniai zemes sklypai) | CC BY 4.0 (attribution VI Registru centras; licence link + change indication mandatory) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/lithuania_terms_20260806.txt) - the dataset record states verbatim that the data may be used and shared without limit under CC BY 4.0 with mandatory attribution, licence link and change indication. RATE-LIMIT TRAP - data.gov.lt serves HTTP 429 aggressively (hit during recon AND build); gpkg_339 honours Retry-After with a 20s floor between requests and must not be made faster. Inner JSON schema not yet measured for the same reason; the loader aborts printing the measured structure on first contact if it is not GeoJSON. |
| Luxembourg cadastre (ACT, Plan Cadastral Numerise) | CC BY 4.0 (attribution Administration du cadastre et de la topographie) | cleared | Cleared 2026-07-31 on the data.public.lu record, licence id cc-by (determination raw/lu_cadastre/). URL-STALENESS TRAP - the resolved download path embeds a reissue timestamp that changes weekly, so a hardcoded URL rots SILENTLY, serving an old edition rather than failing; the stable udata alias is used instead. The published GeoJSON is NOT a FeatureCollection (top level is keyed by layer) and is reprojected to WGS84, so the EPSG:2169 shapefile is used. ACT discloses heterogeneous accuracy, centimetre-level on recent survey down to ~10 m on 1824-derived sheets. Fabric never ownership. |
| Luxembourg INSPIRE electricity transmission network | CC0 1.0 (public domain) | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt, lead 364). data.public.lu dataset record states "license":"https://creativecommons.org/publicdomain/zero/1.0/" - CC0, no attribution obligation, though attribution is carried anyway as house practice (Geoportail Luxembourg / ACT). Fetched directly from the publisher's own download.data.public.lu resource URL (GML, INSPIRE ElectricityNetwork schema), most recent capture dated 2026-06-12 on the publisher's side. |
| Marine Regions EEZ | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Marktstammdatenregister (Bundesnetzagentur, DE) | DL-DE-BY-2.0 (Datenlizenz Deutschland Namensnennung 2.0) | cleared | DL-DE-BY-2.0 - commercial use, redistribution and derivatives permitted with attribution "Marktstammdatenregister der Bundesnetzagentur (dl-de/by-2-0)". Attribution carried on /sources, /licensing and on-row. |
| Maryland parcel boundaries (MD iMAP) | express redistribution grant (MD iMAP) | cleared | Owner names exist only in SDAT's separate LICENSED MdProperty View, which is not used. Mailing-address and site-address columns are published and deliberately NOT read - no owner name is not the same as no personal data. Area carried is the ASSESSED acreage, not the polygon-derived POLYACRES. PAGING TRAP - the service 500s on some offsets and it IS page-size related, so the loader halves the page until it fits rather than retrying at the same size. |
| Meta Relative Wealth Index | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Microsoft Global Building Footprints | ODbL | cleared | ODbL; commercial use permitted; attribution required and share-alike applies to derivative databases. Query results are served as produced works with Microsoft Global ML Building Footprints attribution. |
| MISO (Midcontinent ISO) market reports | MISO market reports terms | derived only | Public market reports, no key; MISO data terms review in progress; fails closed in commercial_safe; on-row licence carried. |
| MITRE ATT&CK for ICS | MITRE ATT&CK Terms of Use (royalty-free, attribution) | cleared | MITRE ATT&CK Terms of Use grant a royalty-free licence for research, development and commercial purposes with the MITRE copyright notice reproduced (verified 2026-07-02). Currently internal-only. |
| NASA FIRMS | US Public Domain | cleared | Public domain; no restrictions. |
| NASA GRACE-FO | US Public Domain | cleared | Public domain; no restrictions. |
| NASA POWER | US Public Domain | cleared | Public domain; no restrictions. |
| National Embedded Capacity Register (ENA / GB DNOs) | Northern Powergrid Open Data Licence v1.0 | cleared | DNO open-data licence modelled on OGL; reuse including commercial with attribution. Combined register published by the DNOs under Ofgem Data Best Practice; per-DNO feeds replace this vehicle when portal registrations land. |
| Natural Earth bathymetry | Public Domain | cleared | Public domain, no restrictions. Attributed on the map surface regardless, as the sibling coastline product is. Same publisher as world_coastline but registered separately, because a licence is a property of the dataset and not of the brand. |
| Natural Earth coastline | Public Domain | cleared | Public domain; no restrictions. |
| NatureScot Protected Areas | OGL v3 | cleared | OGL v3 stated in the dataset''s ArcGIS item metadata (captured 2026-07-18 in raw/naturescot/); attribution "(c) NatureScot, contains Ordnance Survey data (c) Crown copyright". |
| NESO (National Energy System Operator, UK) | NESO Open Data Licence (attribution "Supported by National Energy SO Open Data") | cleared | NESO Open Data Portal licence permits commercial use with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| NESO Electricity Ten Year Statement (ETYS) appendices - GB transmission electrical model | NESO website terms (personal print copy only; commercial exploitation and redistribution forbidden) - NOT the NESO Open Licence | excluded | EXCLUDED = held, never served, and the exclusion is the whole point of the entry. Determination raw/neso_etys/etys_rights_determination_20260810.txt, 2026-08-10. NESO runs TWO instruments over different material and only one of them is open. The NESO Open Licence is OGL-shaped and permits commercial exploitation, but its own scope text binds it to 'datasets as specified within the NESO data portal service', licensed individually; the source `neso` (cleared) covers those. The ETYS APPENDICES are not portal datasets - they are served from /publications/ and /document/<id>/download under the website terms, which permit only 'a print copy ... for your own personal use' and expressly forbid 'otherwise commercially exploit'. The portal's CKAN API was searched rather than assumed - q=ETYS returns exactly one dataset (boundary GEOMETRY, which IS open) while impedance, circuit rating, fault level and nodal demand all return zero. So this data is held for internal use only under Stu's T40 ruling (ingest everything, serve nothing a licence bars) and NO served surface, page, receipt, export or public figure may carry it. Unblocking action is founder ticket ST-23, a written permission request; ETYS is published in discharge of a licence obligation so third parties can model the network, which makes the ask reasonable but does not make it granted. PREFIX TRAP, deliberate - these tables are etys_* and NOT neso_etys_*, because source attribution is FIRST-MATCH by prefix and the cleared `neso` source claims neso_, so a neso_etys_ table would inherit CLEARED status and the gate would fail open; every table is additionally named in source_overrides, which is checked before any prefix. TWO KNOWN PUBLISHER ANOMALIES, preserved verbatim per the no-rewrite rule and recorded here so no future consumer meets them cold - (1) 26 circuits carry zero reactance, and they are explainable - 9 are NGET's own 'Zero Length' circuit type, the rest are sub-kilometre cables and 1-metre OHL, so these are busbar-class connections rather than errors; (2) three circuits publish a winter rating BELOW their summer rating, and one of them looks like a genuine publisher defect - SHE's BLHI41-MOWO41 cable pair is rated 22.7026 MVA in winter against 470 MVA in every other season, a four-decimal value among integers. Any power-flow or headroom use MUST handle that value explicitly - taken at face value it would make a 470 MVA cable look 95% constrained and would manufacture a binding constraint that does not exist. |
| netztransparenz.de (German ÜNB information platform) | netztransparenz WebAPI terms (citation "Quelle - netztransparenz.de") | cleared | Cleared 2026-07-16 after terms review (registered WebAPI publisher client, terms accepted at registration). Source citation "Quelle - netztransparenz.de" carried on /sources, /licensing and in response provenance. Website Impressum consent clause applies to website reproduction; API-delivered data governed by the accepted WebAPI terms. Evidence raw/netztransparenz/impressum_nutzung_20260715.html. |
| New South Wales cadastre (Spatial Services DCS) | CC BY (data.nsw.gov.au b8659c69; Cadastre Web Service package 42c0fdeb states CC BY 3.0 AU) | cleared | Cleared 2026-08-02. NO BULK FILE EXISTS - NSW publishes web services only, so paging the FeatureServer is the intended route. TRAP - the service advertises geoJSON but returns 503 for it, so responses are ESRI JSON parsed by GDAL rather than by hand. TRAP - the publisher EXTENDED its own coded-value domain between the 30 July capture and the 2 August run; read domains from the live service, never from a capture. 8,601 'Dummy' lots are publisher artefacts, kept and labelled so the count reconciles, but arguably should be excluded from any parcel count the screen reports. |
| New York Independent System Operator | NYISO Legal Notice | derived only | NYISO Legal Notice (captured 2026-07-22) is a reserve-all website disclaimer with no express data-use grant and no express prohibition on data use. Derived-only posture; scope of derived context under review. |
| New Zealand primary parcels (Toitu Te Whenua LINZ, LDS layer 50772) | CC BY 4.0 (attribution Toitu Te Whenua LINZ) | cleared | Cleared 2026-08-07 on licence (determination raw/intl_cadastre_recon/new_zealand_terms_20260806.txt) - LINZ states verbatim that data may be used, remixed, distributed and built on 'even commercially' under CC BY 4.0 with credit. TWO CAVEATS RECORDED - (1) property/ownership layers with personal information carry an ADDITIONAL licence; NZ Primary Parcels is a boundary layer but the layer-page badge must be verified on the first keyed session before serving; (2) robots disallows /services to crawlers - the keyed WFS is the LDS-sanctioned programmatic channel and first contact happens under a founder-registered key. |
| NGED Connected Data portal (CKAN) | NGED (WPD) Open Data Licence | cleared | Cleared 2026-07-28 after rights review (founder-directed); the portal's CKAN licence registry declares the 'WPD Open Data Licence' Open-Definition-approved / OKD-compliant (commercial reuse permitted); fetched under registered account. Attribution 'Contains National Grid Electricity Distribution (WPD) open data'. Determination raw/nged/terms_review_20260728.txt; residual = browser capture of the Cloudflare-gated licence text page. |
| Nigeria, Norway and Qatar grid layers (CC-BY batch) | CC BY 4.0 - attribution per table | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt). CC-BY across all three publishers - Nigeria's ArcGIS service states the CC-BY URL in its own copyrightText field, Norway's Zenodo record states cc-by-4.0, Qatar's Opendatasoft record states the CC-BY URL. PAGINATION IS A CORRECTNESS REQUIREMENT HERE, NOT AN OPTIMISATION - the Nigerian FeatureServer returns exactly 1000 features to an unpaginated query, its maxRecordCount, with no flag and no error; the full layer is 5,792, so a loader taking the first response would have stored 17 percent of Nigeria's transmission network and reported success. QATAR IS NOT WHAT ITS TITLE SAYS, found on read - the lead is 'Electricity Sub-Stations by Voltage Level' and it is 8 rows of ANNUAL COUNTS per voltage class, a time series with no geometry and no per-site rows. Held as a fleet-size series and explicitly NOT as a substation layer; third time this week a lead title over-promised its contents. NIGERIA CARRIES ITS PUBLISHER'S OWN FIDELITY CAVEAT - the service description says the lines give 'an indication of extent and general location, but shouldn't be used for precisely locating grid lines', which must travel with any use. PLACE GATE - 15 Nigerian and 1 Norwegian feature sit beyond the national envelope plus a 0.6 degree buffer, consistent with cross-border interconnection and flagged rather than dropped. |
| Nigerian 50-Bus 330 kV power grid dataset | CC BY 4.0 | cleared | CLEARED 2026-08-12. CC-BY-4.0 from the Zenodo record's own metadata (record 12091189). A real POWER-FLOW CASE, not an asset layer - bus type, voltage setpoint, active and reactive DEMAND, active and reactive GENERATION and reactive LIMITS, plus per-branch resistance, reactance and charging susceptance. Taken ahead of larger leads because depth is not capacity - the reactive dimension is the one T74 recorded as absent when Stu named it. EXTRACTED FROM A PDF, which is a risk to manage and not a detail; two guards - the bus count is ASSERTED against the 50 the deposit claims, and every branch endpoint must resolve to a bus in Table 1, so an extraction that loses or invents rows stops rather than loads. ONE HEADER IS AMBIGUOUS AND FLAGGED ON EVERY ROW - Table 2's third numeric column extracts as '12 B (Siemens)', almost certainly a mangled half-B pi-model charging susceptance; the reading is an INFERENCE about a glyph pdftotext could not render and the value is stored exactly as printed. THE HONEST LIMIT - buses carry NAMES and no coordinates, so this is the ETYS case not the Brazil case; real depth, weak stitchability, and nothing here may repair a mapped claim. |
| NIST National Vulnerability Database | US Public Domain | cleared | Public domain; no restrictions. |
| NOAA IBTrACS | US Public Domain | cleared | Public domain; no restrictions. |
| Nord Pool day-ahead area prices (15 Nordic/Baltic areas) | Nord Pool website terms + Market Data Services (Internal Use Licence by default; redistribution is a priced product) | excluded | DETERMINED 2026-08-29 (T143b; evidence raw/tso_terms_20260829/nordpool/ verbatim): "all Market Data products include an Internal Use Licence" and "Redistribution rights need to be purchased separately" - even DERIVED outputs need the priced non-display licence. EXCLUDED from all serving, fail closed; rows stay internal. SEPARATE FLAG FOR STU (ST-46): the website terms also state "Nord Pool does not permit automatic extraction of data", which touches the COLLECTION itself, not just serving - pause-or-license is a founder decision. |
| North Rhine-Westphalia cadastre (Geobasis NRW ALKIS) | Datenlizenz Deutschland Zero 2.0 (dl-de/zero-2-0); NO attribution condition | cleared | Cleared 2026-08-02. Licence is ZERO not BY, confirmed at THREE independent points including the CSW record for the bulk-download product specifically, which establishes it covers the files and not only the web services. AXIS TRAP AVOIDED BY ROUTE CHOICE - the NRW WFS publishes EPSG:4258 latitude before longitude; the bulk GeoPackages are EPSG:25832 projected. The published Lagebezeichnung (street and house number) is deliberately NOT read. |
| Northern Powergrid Open Data | Northern Powergrid Open Data Licence (OGL-style) | cleared | DNO open-data licence modelled on OGL; reuse including commercial with attribution (portal terms verified 2026-07-11; re-verify licence text at first commercial use). |
| Norwegian cadastre map (Kartverket, Matrikkelen - Eiendomskart Teig) | CC BY 4.0 (attribution Kartverket) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/norway_terms_20260806.txt) - the catalogue records CC BY 4.0 with 'Ingen begrensninger pa bruk er oppgitt' and the WFS purpose field states the information may be used by anyone under Utleveringsforskriften par.3(2), the statutory carve-out that lets geometry+ID flow openly while owner data stays permissioned. Resolved the 'NLOD or CC BY' lead to CC BY 4.0. National file exists ONLY as FGDB/PostGIS in EPSG:25833 (format codelist measured 2026-08-07), artefact-gated in gpkg_340. |
| NSIP register (Planning Inspectorate) | OGL v3 (Crown body publication) | under review | Crown body publication expected under OGL v3 but the download endpoint does not state terms explicitly — under review until confirmed. Fails closed in commercial_safe mode. |
| NZ Electricity Authority - Transpower PSS/E power-flow cases | no express licence found - bare copyright assertion, terms URLs 404 | excluded | INTERNAL_ONLY 2026-08-12 (determination raw/nz_ea_psse/nz_ea_psse_rights_determination_20260812.txt). The EA's PSS/E v32.1 solved cases - North and South Island, light and peak, 100 MVA / 50 Hz base, exported from Transpower's PowerFactory Central Data Set - carrying bus voltage AND ANGLE, loads, generators with reactive limits, and branches with impedance and three rating sets. WHY BARRED - both EA terms URLs 404 and the site asserts bare copyright, so there is no express licence, and absence of a permissive licence is a bar not a gap, exactly as ruled for ETYS and TYTFS. WHY IT IS THE STRONGEST CANDIDATE YET TO CLEAR - the release notes shipped WITH the download state the System Operator is required to submit these files 'for the Industry to use', a regulatory obligation to publish for use, where NESO's ETYS terms actively FORBADE commercial exploitation. Purpose is still not permission. NZGOAL would clear it outright if it applies but was NOT found stated for this dataset and is not relied on. THE DOWNLOAD IS WHY THE POSITION IS KNOWN - a page-only read left it unresolvable and the decisive sentence was inside the release notes, which is the general lesson that a dataset often carries its own terms. PREFIX DELIBERATELY FAR FROM THE CLEARED SOURCE - nz_psse_* and never nz_grid_*, because attribution is first-match by prefix and one country now holds two grid sources under two different licences. |
| OECD-NEA UDEPO | IAEA Nucleus Terms (attribution) | under review | IAEA database (UThDEPO). IAEA Nucleus terms permit use in commercial products with acknowledgment (verified 2026-07-02); confirming no database-specific restriction applies. |
| OFAC SDN List | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). |
| ONS Brazil - Operador Nacional do Sistema Eletrico (transmission network model) | CC BY 4.0 (Creative Commons Atribuicao) - attribution "ONS - Operador Nacional do Sistema Eletrico" | cleared | CLEARED 2026-08-11 (determination raw/ons_brazil/ons_rights_determination_20260811.txt). Licence read from the PUBLISHER'S OWN CKAN API, not from a third party's tag - all three datasets return license_id cc-by, license_title 'Creative Commons Atribuicao', and the portal states it guarantees 'acesso, reuso e redistribuicao'. NO site-wide terms page exists (/pages/termos-de-uso verified 404 and the body captured), recorded rather than glossed because the NESO case three days earlier turned on the OPPOSITE configuration - an open portal licence sitting beside restrictive website terms with the valuable technical data on the restricted side. Here there is no competing instrument. ATTRIBUTION REQUIRED - 'ONS - Operador Nacional do Sistema Eletrico' - and CC-BY carries NO share-alike, unlike the ODbL governing pypsa_eur_base, so the two must never share one house attribution string. WHAT IT CONTAINS is a full transmission model, the first outside Europe this estate may both serve and traverse - substations with coordinates, lines with resistance/reactance/shunt and long and short-term operating capacity, transformers with three winding voltages and MVA, and num_barra, the operator's own published BUS number, on every table. FORMAT TRAPS ESTABLISHED BY INSPECTION - UTF-8 (probed, since a Brazilian feed is a classic place to find latin-1), SEMICOLON delimiter, and a decimal POINT rather than the Brazilian comma, so no locale conversion is applied; applying one would have corrupted every numeric value while leaving the row count identical. Text columns are space-padded to fixed width and the padding is stripped, a stated decision and not a value rewrite. |
| ONS Census 2021 (E&W) population and household estimates by unit postcode | OGL v3, Crown copyright; attribution Source Office for National Statistics licensed under the Open Government Licence v.3.0 | cleared | Cleared 2026-08-16 (determination raw/ons_census2021_postcode/ons_census2021_postcode_rights_determination_20260816.txt). Never treat as a postcode directory; join via ONSPD epochs or UPRN to OA instead. |
| ONS Census 2021 (England and Wales) topic-summary tables at Output Area grain | OGL v3, Crown copyright; attribution Source Office for National Statistics licensed under the Open Government Licence v.3.0 | cleared | Cleared 2026-08-16 (determination raw/nomis_census2021_bulk/nomis_census2021_bulk_rights_determination_20260816.txt). OGL v3 permits commercial exploitation with attribution. 12 of ~77 TS bulk zips taken (priority topics); NI not captured (different statistical body). |
| ONS Open Geography | OGL v3 | cleared | ONS Open Geography portal boundary products are OGL v3; attribution "Source - Office for National Statistics licensed under the Open Government Licence v.3.0; contains OS data (c) Crown copyright". |
| ONS Open Geography OA 2021 boundaries (E&W) and OA-LSOA-MSOA-LAD22 lookup | OGL v3; attribution Source Office for National Statistics licensed under the Open Government Licence v.3.0; Contains OS data (c) Crown copyright and database right 2026 | cleared | Cleared 2026-08-16 (determination raw/ons_open_geography/ons_open_geography_rights_determination_20260816.txt). BGC generalised variant chosen deliberately (stitching use, not cartographic precision); LSOA/MSOA boundary files and PW centroids deliberately not taken. |
| ONS Postcode Directory (May 2026) | OGL v3 for GB; attribution Source Office for National Statistics licensed under the Open Government Licence v.3.0; Contains OS data (c) Crown copyright and database right 2026; Contains Royal Mail data (c) Royal Mail copyright and database right 2026. Northern Ireland data NOT under OGL (separate LPS licence) | cleared | Cleared for GB 2026-08-16 with NI carve-out (determination raw/ons_onspd/ons_onspd_rights_determination_20260816.txt). Commercial use of NI data needs a separate Land and Property Services licence, so NI rows are excluded from the loaded table entirely. |
| ONS UPRN Directory (Dec 2025, Epoch 123) and National Statistics UPRN Lookup (Jun 2026, Epoch 127) | OGL v3; attribution Source Office for National Statistics licensed under the Open Government Licence v.3.0; Contains OS data (c) Crown copyright and database right 2026; pcds field additionally Contains Royal Mail data (c) Royal Mail copyright and database right 2026 | cleared | Cleared 2026-08-16 (determination raw/ons_open_geography/ons_open_geography_rights_determination_20260816.txt). Both epochs kept deliberately; the uprn_oa stitch prefers NSUL where they disagree and states its basis per row. |
| OpenAQ | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| OpenFlights | ODbL | cleared | ODbL; commercial use permitted; attribution required and share-alike applies to derivative databases. Query results are served as produced works with OpenFlights contributors attribution. |
| OpenInfraMap | ODbL | cleared | ODbL; commercial use permitted; attribution required and share-alike applies to derivative databases. Query results are served as produced works with attribution (© OpenStreetMap contributors; OpenInfraMap is OSM-derived). |
| OpenSanctions | CC BY-NC 4.0 (commercial licence available) | under review | No raw records are served. OpenSanctions feeds internal analytics only. Publisher offers a commercial licence, which is the planned route (verified 2026-07-02). |
| OpenSky Network ADS-B aircraft positions | OpenSky General Terms of Use & Data License Agreement | under review | Terms reviewed 2026-07-18 (archived at review date) - the OpenSky licence covers non-profit research and education only; commercial use and any operational or automated API use require a written licence from OpenSky Network. A licence decision is pending; no aircraft-derived data serves in commercial-safe mode and none enters any public or anonymous surface. |
| OpenStreetMap | ODbL 1.0 | cleared | ODbL; commercial use permitted with attribution and share-alike on derived databases. Attribution carried on /sources. |
| OpenStreetMap planet power extract | ODbL | cleared | ODbL; commercial use permitted; attribution required and share-alike applies to derivative databases. Query results are served as produced works with attribution (© OpenStreetMap contributors). Ruled cleared by Stu 2026-08-18. ACQUISITION PROVENANCE, not the rights basis - the planet power extract was purchased for £700 in March 2024, but ODbL governs use regardless of how the copy was obtained, so the purchase evidences provenance while the licence carries the obligations. This field was ABSENT until 2026-08-18 and the tier gate is an allowlist, so osm_power_* (~165M rows, 17GB of grid topology) fail-closed out of the API, MCP and the public surface the whole time. |
| OS Open UPRN | OS OpenData licence (OGL-aligned); attribution Contains OS data (c) Crown copyright and database right 2026 | cleared | Cleared 2026-08-15 (determination raw/os_open_uprn/determination_20260815.md); licence.txt shipped inside the archive. Keyless download via the OS Downloads API. |
| OurAirports | Public Domain | cleared | Public domain; no restrictions. |
| Overture Maps buildings (OSM + Microsoft + Google Open Buildings + Esri) | ODbL 1.0; attribution (c) OpenStreetMap contributors, Overture Maps Foundation, plus per-source credits | cleared | Cleared 2026-08-15 (determination raw/overture_buildings/determination_20260815.md). ODbL share-alike ATTACHES to the derived parcel_building_stats table (same posture as the PyPSA T90 correction) - the served layer is offered under ODbL terms with full attribution. Chosen over re-streaming raw MS footprints (T110 substrate decision) - superset corpus, height/floors/class attributes, stable GERS ids. |
| Packet Clearing House IXP directory | CC BY-NC-SA 3.0 | under review | CC BY-NC-SA 3.0 (publisher-stated, verified 2026-07-02). Only per-country counts are exposed; NC clause being resolved with publisher. |
| ParcelMap BC (British Columbia parcel fabric) | Open Government Licence - British Columbia v2.0 | cleared | Cleared 2026-07-31 under OGL-BC v2.0, which grants a worldwide, royalty-free, perpetual, non-exclusive licence including for commercial purposes (determination raw/ca_bc/). Attribution is the licence fallback wording since the catalogue specifies no custom string - 'Contains information licensed under the Open Government Licence - British Columbia'. NO OFFICIAL MARKS and NON-ENDORSEMENT conditions apply. TWO CATALOGUE RECORDS EXIST AND ONLY ONE IS USABLE - the fully-attributed variant carrying owner names is 'Access Only', PROTECTED A, Government audience, and must never be fetched; here fabric-not-ownership is a licence boundary, not a house rule. PUBLISHER TRAP - the two bulk files share the title 'Parcel Fabric File Geodatabase' but are NOT the same data (3005 holds 2,491,576 features, 3153 only 1,730,818); the 3005 product is used. Not the authoritative legal boundary, which is the plan of survey. Federal land and First Nations reserves are thinly represented; strata units share the parent parcel geometry. |
| PeeringDB | CC BY-SA 4.0 | cleared | CC BY-SA 4.0; commercial use permitted; attribution and share-alike on adaptations. Attribution carried on /sources, /licensing and in response provenance. |
| Per-jurisdiction transmission voltage ceilings (methodology reference) | House-curated statements of public regulatory fact; each row's basis names the TSO source verbatim | cleared | SERVED ON STU'S RULING 2026-08-29 ("serve it") - the last table outside the governed surface, closing T128 at 647/647. House-authored curation over published TSO facts; the basis column carries the citation per row, so serving it exposes the methodology's own evidence rather than any third-party database. |
| PJM Interconnection | PJM Data Miner 2 Terms of Use | under review | Data Miner 2 Terms of Use permit commercial use including publishing and making derivatives; republishing verbatim data is reserved to PJM Members. Derived-only posture; scope of derived context under review. |
| planning.data.gov.uk designations (England) | OGL v3 | cleared | OGL v3; attribution carried on this register and /licensing. The platform aggregates authoritative publishers (MHCLG collections, Natural England). Recorded 2026-07-17. |
| Polish cadastre aggregation (GUGiK EGiB dzialki ewidencyjne) | statutory free access (Prawo geodezyjne art. 40a) for parcel identifiers + geometry; attribution GUGiK | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/poland_terms_20260806.txt) - statutory basis art. 40a ust. 2 pkt 1 lit. i quoted verbatim in the capture; the national WFS GetCapabilities states Fees none and AccessConstraints none. ROUTE - the country GeoParquet package from the Pobierz-dane module is the preferred artefact but its URL is built client-side in the SPA and could not be pinned from this box (bounded ~30 min attempt logged in T57); one founder browser minute pins it into PL_EGIB_PARQUET_URL and gpkg_345 switches routes automatically; until then the verified WFS UslugaZbiorcza STARTINDEX paging route stands. The KIEG WMS GetFeatureInfo harvesting ban does NOT touch the WFS, which exists to aggregate the powiat services for bulk retrieval. |
| Public procurement notices (multi-jurisdiction) | OGL v3 / EU reuse / US Public Domain / OGL-Canada (per row) | cleared | All constituent sources are open-licensed for reuse including commercial (OGL v3 UK; EU reuse Decision 2011/833/EU with acknowledgment; US public domain; OGL-Canada; World Bank CC BY 4.0; BOAMP Licence Ouverte Etalab; AusTender CC BY 3.0 AU incl. the ATM RSS; NZ GETS under NZGOAL/CC BY 4.0 per the 2026-08-18 OECD recon, endpoint tested from this box; Colombia SECOP II via datos.gov.co SODA under CC BY-SA 4.0 attribution CCE, share-alike noted, determination 2026-08-22 - the OCDS channel declares ODC-PDDL in-band and is the bulk route; Netherlands TenderNed TNS API under CC0 1.0 declared by both the API spec and data.overheid.nl, determination 2026-08-23; Spain PLACSP syndication under reproduction-with-attribution per the aviso legal + Ley 37/2007, determination 2026-08-23 with the robots-vs-documented-channel conflict recorded in the LICENCES comment and discovery notes; Germany Bekanntmachungsservice under CC0 verbatim from its Open-Data-Richtlinie with a measured 32 percent national-only share, determination 2026-08-23; Norway Doffin under CC BY 4.0 per the data.norge.no catalog record, publisher DFO, determination 2026-08-23; Denmark udbud.dk UNDER_REVIEW - Danish CC0 PSI default + sibling-dataset CC0 but no on-platform licence text, servable-with-caution pending KFST confirmation; Finland Hilma SHIPPED 2026-08-23 via the licensed avp-read eForms read API (subscription created by Stu under ST-40, lukurajapinta terms accepted - field-of-use grant, revocation clause recorded; key in .dno_portals.md); harvest walks the eNotice id sequence backward from a binary-searched head, parses full eForms XML, and links the verified public/procedure/{p}/enotice/{n} route read from the site router bundle - the naive notice-id URL mapping was PROVEN WRONG (site and eForms id spaces are disjoint); Sweden closed as VERIFIED NONE - a registration regime with five commercial operators and no open channel; BATCH 3 WAVE 1 2026-08-24: Portugal BASE/IMPIC via the dados.gov.pt yearly anuncios bulk, public domain per the catalog licence field other-pd, full national tier incl. ajuste direto, awards live in the sibling contratos files and are not yet harvested; Greece KIMDIS Open Data API under CC BY 4.0 quoted verbatim from its help page, 350 req/min stated limit, notices carry no CPV at top level so the el_energy text family is the scoring channel; Estonia RHR monthly eForms-UBL bulk under CC BY-SA 3.0 EE with the SHARE-ALIKE flag noted, unpadded-month path quirk measured; Latvia IUB daily notice JSON under CC0 1.0 with eForms-BT-mapped fields; BATCH 3 WAVE 2 2026-08-24: Poland e-Zamowienia BZP under CC0 per dane.gov.pl dataset 916 published by UZP itself, harvested as an ascending PublicationDateFrom cursor walk after every paging param proved silently ignored; Austria USP Kerndaten central list under CC BY 4.0 per the KDQ feed records, ANKOe CC-BY-ND document sets deliberately not taken; Ireland eTenders register CSV under CC BY 4.0 from the Office of Government Procurement, window anchored on the file's own head because the register lags the live platform by about two months; Czechia VVZ live register UNDER_REVIEW - no licence instrument but MMR's own catalog declares no copyright works and no database right on its exports; Slovakia UVO Vestnik bulletins via data.slovensko.sk under machine-declared CC BY 4.0 on all three axes, harvested only from the open-data copies because uvo.gov.sk robots disallows crawling, row raw keeps selected eForms BT leaves only because the full trees carry personal contact data). Attribution carried per row in src_attr. |
| PUDL (Public Utility Data Liberation, Catalyst Cooperative) | CC BY 4.0 (data; code MIT) | cleared | CC BY 4.0 with attribution to Catalyst Cooperative; underlying FERC/EIA content is US public domain. |
| PyPSA-Eur prebuilt OSM electricity network | ODbL 1.0 (OpenStreetMap-derived) | cleared | Cleared 2026-07-29 after rights review; Zenodo record metadata declares odc-odbl (ODbL 1.0), same posture as the OSM estate. Commercial use permitted; attribution (PyPSA-Eur Zenodo 18619025 + OpenStreetMap contributors) carried. SHARE-ALIKE NOW BINDS IN A WAY IT DID NOT ON 07-29, and this note is updated rather than left describing a use that has changed - the 07-29 reasoning was that use was an internal modelling substrate so redistribution did not arise. As of 2026-08-11 (gpkg_364) the network itself is loaded and SERVED, and serving these rows is redistribution of a derivative database, so the ODbL share-alike condition attaches to the served layer - anyone taking the database onward takes it under ODbL. Attribution is required at BOTH levels, PyPSA-Eur and OpenStreetMap upstream. Determination raw/pypsa_eur_base/terms_review_20260728.txt |
| Queensland cadastre (Department of Resources) | CC BY 4.0 (attribution State of Queensland, Department of Resources) | cleared | Cleared 2026-08-02. LAYER TRAP - the obvious layer 1 'Land Parcels' declares supportsPagination FALSE and cannot be bulk-read; layer 4 'Cadastral parcels' is the paginable one. OBJECTID exists but is NOT declared in the fields array and the layer reports objectIdField null, so key-based paging needs returnIdsOnly to discover it. Positional accuracy IS published per parcel as acc_code and is NOT carried, because the vehicle has nowhere faithful to hold it. |
| RE100 Members | Climate Group terms (proprietary) | excluded | Climate Group site terms require a licence for commercial use (verified 2026-07-02). Members list not exposed on any commercial surface; internal evaluation only. |
| Registers of Scotland INSPIRE Cadastral Parcels | OGL v3 (Contains Registers of Scotland data (c) Crown copyright) | cleared | Cleared 2026-07-29; the download service states OGL v3 with click-through acceptance (capture raw/ros_scotland/mapcloud_index_20260729.html, determination terms_review_20260729.txt). Attribution; Contains Registers of Scotland data (c) Crown copyright, licensed under the Open Government Licence v3.0. |
| Renewable Energy Planning Database (DESNZ) | OGL v3 | cleared | Open Government Licence v3; commercial use permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| RPA Parcel Points (England) | OGL v3 | cleared | CLEARED 2026-08-13 (batch_determinations_20260813.txt group 4 EA_RPA, lead 4502). OGL v3 read from environment.data.gov.uk's own footer; publisher's CKAN package_show record additionally states "There are no public access constraints to this data." THE LEAD'S ?download=true URL IS A JS-RENDERED NEXT.JS PAGE WITH NO STATIC DOWNLOAD LINK - the real distributions live on the CKAN package_show record (WMS, WFS, OGC API - Features), captured. Output format was SNIFFED not assumed - a live limit=100000 request against the OGC API Features endpoint before writing the fetch loop confirmed paginated GeoJSON with a numberMatched total and documented startIndex/limit rel=next paging. No reprojection needed, CRS84 (lon/lat) already. Size ~low hundreds of MB across the full paginated fetch, well under the 2GB ceiling - no boundary hit. |
| Scotland Heat Map - geothermal potential | OGL v3 | cleared | CLEARED 2026-08-13 (batch_determinations_20260813.txt group 2 GOV_SCOT_HEAT, leads 4736,4737,4738,4769,4771,4772,4773,4776,4783,4784,4785,4806 - all 12 resolving to the same landing page). OGL v3 read from the landing page's own footer, cross-checked at 4736/4737/ 4738. THE LANDING PAGE ITSELF HAS NO DIRECT DOWNLOAD LINKS - its "Detailed Scotland Heat Map data" sub-page states the full detailed dataset requires a data-sharing agreement ("shared with local authorities and other organisations through data-sharing agreements"), NOT an open download. The "Access to detailed data for other organisations" sub-page names the actually-open route BEFORE any agreement request - the Scotland Heat Map Interactive website "and its associated downloads and web mapping services" at heatmap.data.gov.scot, an Astun/ iShare OGC stack with no login. Layer list pulled from the site's own map config (getdata.aspx?type=MapSource&RequestType=JSON&ms=mapsources/heatmap) and confirmed against its WFS/WMS GetCapabilities, both captured. TLS NOTE - the site's certificate expired 2026-08-12, one day before this capture; genuine server-side expiry, fetched with GDAL_HTTP_UNSAFESSL and content verified structurally against the publisher's own JSON/XML, not blind-trusted. RASTER SPLIT - the map config's own "type" field marks the Heat Demand and Confidence grids (hmp_*/conf_* at 50m/250m/500m/1km) as type=base (WMS/tile only, absent from the WFS featuretype list, /geotiff/{layer} export 404s directly); the 50m grid alone spans tens of millions of cells across Scotland - captured-not-tabled with reason, no scot_heat_demand table exists from this loader. The three geothermal layers are type=ows, real GeoServer/MapServer WFS featuretypes, EPSG:27700, confirmed via live GetFeature before writing the loader. DELIBERATELY SKIPPED per house rule - Social_Rent_Datazone_1 ("Data Zone (2011) % socially renting"), the Socially Renting Households layer, out of scope for this pass though it IS a real ows/WFS vector layer (104MB GML, 6,976 data-zone polygons); captured to raw/gov_scot_heat/ anyway so a future pass need not re-fetch it. |
| Scotland's Census 2022 (NRS) topic tables at Output Area grain | OGL v3, Copyright National Records of Scotland | cleared | Cleared 2026-08-16 (determination raw/nrs_census2022/nrs_census2022_rights_determination_20260816.txt). NRS-linked copyright page 404s; OGL position evidenced from the live NRS footer, captures retained. 59 further UV tables and all 47 MV tables retained raw, not loaded. |
| Scottish EPC Register domestic extract (Scottish Government) | OGL v3 (address fields under Royal Mail copyright notice) | cleared | Cleared 2026-08-15, CONFIRMED from inside the download 2026-08-16 - the embedded publication note states all fields other than ADDRESS1/ADDRESS2/POST_TOWN/POSTCODE are OGL. Address fields held, never served raw. statistics.gov.scot is RELAY-ONLY from this estate (primary egress connection-dropped). Publisher caveat that travels - the extract is valid current-format lodgements only; a missing property is not an assertion of no certificate. |
| SEC EDGAR | US Public Domain | cleared | Public domain; no restrictions. |
| Secretaria de Energia Argentina - high voltage transport network | CC BY 4.0 - attribution "Secretaria de Energia, Argentina" | cleared | CLEARED 2026-08-12 (determination and captures in raw/ar_energia/). CC-BY-4.0 read from the publisher's own CKAN license_id across all four transport datasets. Attribution to the Secretaria de Energia, Argentina; no share-alike. REACHABLE ONLY THROUGH THE US RELAY, recorded as a property of the SOURCE and not a footnote about our plumbing - a direct request from this box times out with no response at all, while the same request through kortex-relay returns 200. A source that looks dead from one egress and healthy from another is not dead, and the earlier global sweep would have written Argentina off on that timeout; any re-fetch MUST use the relay. WHAT ARRIVED - 1,299 high-voltage circuits with design voltage, conductor section, material, owner and concession, and 323 transformer substations of which 307 carry INSTALLED CAPACITY plus entry and exit voltages. Capacity is the dimension this estate is thinnest on worldwide, so this is worth more than Argentina's capacity ranking suggests. TWO PUBLISHER AMBIGUITIES PRESERVED RATHER THAN RESOLVED - the capacity column is spelled potencia_instalada_mv, and 'mv' is not a power unit; it is almost certainly MVA and it is NOT renamed, because asserting the unit would be our claim wearing the operator's authority. And tension_salida is multi-valued ('13.2 | 33'), stored as published text rather than split, since splitting would invent rows the publisher never asserted. ENCODING TRAP - both files are UTF-8 WITH BOM; read as plain utf-8 the first header becomes a mojibake 'nombre' and every lookup on it returns None while the row count stays perfect, a silent whole-column loss. Read with utf-8-sig and guarded by an assertion. |
| Sentinel-5P TROPOMI | Copernicus open licence | cleared | Copernicus licence; free use including commercial, with attribution. |
| Singapore cadastre (SLA Cadastral Land Parcel via data.gov.sg) | Singapore Open Data Licence (attribution Singapore Land Authority) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/singapore_terms_20260806.txt) - the Open Data Licence grants a worldwide, perpetual, royalty-free licence to use, copy, distribute, modify and adapt the datasets 'whether commercially or non-commercially'; no personal-data rights granted, no endorsement implication. ACCESS NUANCE - the metadata API 403s without a key but the initiate-download endpoint the portal front end itself uses answers anonymously (verified 2026-08-07); gpkg_347. |
| SIPRI Military Expenditure | SIPRI terms (commercial use requires authorisation) | under review | SIPRI terms require advance authorisation (normally a royalty) for commercial use (verified 2026-07-02). Country military-expenditure values are exposed on Jurisdiction records; pursuing authorisation or replacement with derived classes. |
| Slovenian cadastre (GURS, Kataster nepremicnin) | CC BY 4.0 (attribution Geodetska uprava RS) | cleared | Cleared 2026-08-07 (determination raw/intl_cadastre_recon/slovenia_terms_20260806.txt) - OPSI records CC BY 4.0 and portal Prostor grants free use citing source and authorship. ROUTE - ipi.eprostor.gov.si robots.txt disallows /wfs*/ except GetCapabilities and the old eGP host is dead, so bulk is the sanctioned channel; the JGP service API (no robots restriction, public OpenAPI) mints the package download token ANONYMOUSLY on a bare GET (group 122 composite product 25, filter DRZAVA=1; verified end-to-end 2026-08-07, correcting the same-day earlier read that a SPA session was needed), so gpkg_342 self-fetches at load time with a local artefact as fallback. CRS D96/TM EPSG:3794 artefact-gated. |
| Smithsonian Global Volcanism Program | Smithsonian terms (non-commercial; PD nuance noted) | excluded | GVP site terms are non-commercial (verified 2026-07-02; underlying compilation is US federal work; public-domain status noted, may be cleared later). Not exposed on any commercial surface. |
| Southern Company Wholesale | Southern Company website terms (proprietary) | excluded | Southern Company site terms restrict reuse (verified 2026-07-02). Prices retained for internal calibration only; never served, and feed no downstream product. |
| SP Energy Networks Open Data | SPEN Open Data Licence (portal terms) | under review | Portal licence terms to be verified for commercial redistribution — under review; fails closed in commercial_safe mode. The SPM heatmap dataset metadata declares CC BY 4.0 verbatim (captured raw/spen_opendata/spm_capacity_heatmap_dataset_metadata_20260821.json) — a portal metadata claim awaiting a human determination; on a cleared ruling the flip is a registry update, never a reload. |
| Spain CNIG BTN100 Theme 7 energy layers | CC BY 4.0 - prescribed attribution string "BTN100 2015-2021 CC-BY 4.0 scne.es" | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt, lead 400) on the publisher's own PRESCRIBED attribution string, read verbatim from CNIG's licence popup - "BTN100 2015-2021 CC-BY 4.0 scne.es" (Tipo 1 of three permitted forms), the same prescribed-string convention recorded for Northern Powergrid; not paraphrasable. FETCH SHAPE FOUND 2026-08-13 - CNIG's Centro de Descargas is a stateful Java/Tomcat portal (JSESSIONID, ajax forms returning HTML), the same portal-not-API class as Opendatasoft and data-fair, but unlike ENW/EDF-SEI it does not gate the file behind a login - a three-step plain-POST sequence (archivosSerie to list files for all of Spain -> initDescargaDir to get a download ticket -> descargaDir to stream the ZIP) works with no browser automation and no session cookie carried between steps. Theme 7 has exactly ONE vintage for all of Spain (2015 - unlike Theme 6/transport which has 2015 and 2021, energy was not re-surveyed in the 2021 pass), so 2015 is the current publisher vintage, not a stale choice. THREE LAYERS, THREE TABLES - es_transmission_lines (LineString, 2,460), es_power_stations (Polygon, 2,558), es_substations (Point, 2,578); TIPO_0702 line-type code kept verbatim, undecoded (no legend shipped in this ZIP). CRS PRESERVED AS PUBLISHED - source .prj is ETRS89 geographic (EPSG:4258), stored as SRID 4258 rather than relabelled 4326 (TOWGS84 shift is all-zero, so numerically coincident, but the SRID itself is a fact kept verbatim). |
| Spanish cadastre (Direccion General del Catastro, INSPIRE) | DGC INSPIRE licence - transformed use only | derived only | The product-specific INSPIRE licence permits commercial use of TRANSFORMED products and expressly forbids disseminating the original supplied information by any medium (determination raw/es_catastro/terms_review_20260730.txt). Raw parcel geometry is never served; the screen returns counts and area shares only, which is a transformed work. Attribution; Direccion General del Catastro (Ministerio de Hacienda). |
| SPP (Southwest Power Pool) Marketplace | SPP Marketplace terms | cleared | Public Marketplace file API, no key; SPP terms review in progress; fails closed in commercial_safe; on-row licence carried. |
| SSEN Distribution Open Data | CC-BY-4.0 | cleared | CC-BY-4.0 stated in the portal's CKAN metadata per dataset (captured 2026-07-18); reuse including commercial with attribution "SSEN Distribution". |
| Swedish property division (Lantmateriet Fastighetsindelning via Geotorget) | CC BY 4.0 under HVD villkor (attribution product name + (c)Lantmateriet + processing note) | cleared | Cleared 2026-08-07 on licence with an access caveat (determination raw/intl_cadastre_recon/sweden_terms_20260806.txt) - the HVD fee wall fell 2025-02-03 and villkor LM2025/009266 3.1 grants CC BY 4.0 verbatim including commercial contexts, with attribution of product name, (c)Lantmateriet, processing note and licence notice. UNRESOLVED AND RECORDED PER DISCIPLINE - whether this product falls under the plain HVD villkor or the personal-data villkor (approved-purpose decision, LM2025/009269 3.3) could not be established without an account; pure parcel geometry points to the plain villkor but Lantmateriet does not name the provning products. Resolve at registration BEFORE any load; treat the PU gate as possible until disproven. |
| Swiss amtliche Vermessung parcels (geodienste.ch av_0 aggregate) | per-canton free use, attribution mandatory in 14 of 21 free publishers | cleared | Cleared 2026-08-07 for the free cantons (determination raw/intl_cadastre_recon/switzerland_terms_20260806.txt) - services.json opendata terms verbatim 'Freie Nutzung. Quellenangabe ist Pflicht.' for 14 publishers, 'Freie Nutzung' for AR GL SH SO ZH, 'keine Angabe' for GR SG whose canton terms pages govern (URLs in the capture). Switzerland is NOT EU - no HVD regime; availability is cantonal policy and can change per canton. FAIR USE IS A LOAD-BEARING CONSTRAINT - 10 WFS requests/min and an explicit no-high-intensity-bots clause, so gpkg_343 paces at 6.5s/request and an overnight runtime is the price of compliance. INTERLIS bulk zips rejected (no converter on the box); WFS carries the same RESF layer in EPSG:2056. |
| Tasmania + ACT parcel fabrics (theLIST, ACTmapi) | Tasmania CC BY 3.0 AU (the LIST (c) State of Tasmania); ACT CC BY 4.0 ((c) Australian Capital Territory) | cleared | Cleared 2026-08-07 (determinations raw/intl_cadastre_recon/tasmania_terms_20260806.txt + act_terms_20260806.txt) - Tasmania under CC BY 3.0 AU with verbatim 'even commercially'; ACT under item licenseInfo 'Creative Commons By Attribution 4.0 (Australian Capital Territory)'. ACT LAST_UPDATE arrives as ArcGIS epoch milliseconds and is converted to an ISO date - an encoding conversion, not a value rewrite. gpkg_346 (the gpkg_334 idiom re-implemented; gpkg_334 itself untouched while tranche 2 ran). |
| Telecom site layer (OpenInfraMap + PeeringDB) | per row (ODbL / PeeringDB terms) | cleared | Constituents cleared; per-row source discloses which licence applies. |
| TeleGeography Submarine Cable Map | CC BY-NC-SA 3.0 (historic; current terms unconfirmed) | under review | Cable geometries are currently in the graph. Historic publication was CC BY-NC-SA 3.0; current terms could not be confirmed on live pages (2026-07-02). Seeking publisher licence or alternative source. |
| Transparency International CPI | CC BY-ND 4.0 | under review | CC BY-ND; redistribution with attribution permitted; the NoDerivatives clause is under review for Kortex-derived composite scores. |
| Transpower New Zealand - national grid asset layer | CC BY 4.0 - attribution "Transpower New Zealand Limited" | cleared | CLEARED 2026-08-12 (determination raw/transpower_nz/transpower_rights_determination_20260812.txt). CC BY 4.0 read from the publisher's own MACHINE-READABLE DCAT-US catalogue feed, the strongest evidence short of a signed grant - all ten catalogue datasets carry the same licence URL. Attribution to Transpower New Zealand Limited; NO share-alike, so it must never share an attribution string with the ODbL-governed pypsa_eur_base. WHAT IT IS NOT, stated because the register must not overclaim - this is an ASSET layer, not an electrical model. There are no impedances and no ratings, so has_network_model stays FALSE for NZL on the strength of it. New Zealand's actual model is published SEPARATELY by the Electricity Authority as PSS/E .raw cases exported from Transpower's PowerFactory files, the most open grid publication of any jurisdiction surveyed, and that is a separate determination not yet done. THE JOIN KEY IS PUBLISHED, which is why the layer is worth taking - MXLOCATION on a circuit encodes its endpoint site codes (AHA-DOB-A is Arahura to Dobson) and those are the Sites layer's own codes, so line-to-site topology is the operator's assertion rather than a proximity guess; 168 of 230 circuits resolved both endpoints and the other 62 were left NULL rather than guessed. DELIBERATELY SKIPPED - the catalogue's 'Line Area Contacts' layer, a personal data risk with no siting value, and the skip is recorded because what was not taken matters as much as what was. |
| TSO market feeds - cleared-publisher serving views | Per underlying publisher (all members cleared; attribution strings per publisher in their own entries) | cleared | T143b RESOLUTION 2026-08-29 - the registry-substrate unlock for the seven-week tso serving hold: per-publisher determinations landed, and the cleared subset serves via views (a registry entry, not new code). Held out by design - Nord Pool areas (excluded, internal-use licence), AESO (excluded pending written permission, ST-45), MISO/ISO-NE/NYISO (derived_only - analytics serve, raw rows never). IESO reproductions must carry its prescribed verbatim notice. |
| UCDP / PRIO | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| UK OFSI Consolidated List | Open Government Licence v3 | cleared | OGL v3; commercial use permitted with attribution (verified 2026-07-11). |
| UK Power Networks Open Data | UKPN Open Data Licence (portal terms) | under review | Portal licence terms to be verified for commercial redistribution — under review; fails closed in commercial_safe mode. |
| UK UPRN-anchored join spine (derived stitch tables) | Composite of parents; every served derivative carries Source Office for National Statistics licensed under the Open Government Licence v.3.0 plus Contains OS data (c) Crown copyright and database right 2026; ppd_postcode_oa additionally Contains Royal Mail data (c) Royal Mail copyright and database right 2026 | cleared | Derived exclusively from cleared parents (onsud_nsul, ons_onspd, os_open_uprn, hmlr_ppd, epc_scotland, land_inspire); attribution conditions of the parents attach to every serve. Built 2026-08-16 by gpkg_412/gpkg_413. |
| UN Comtrade | UN Comtrade terms | under review | Bulk redistribution requires a UN Comtrade licence; evaluating derived-only posture. |
| UN General Assembly voting (Voeten) | CC0 | cleared | CC0; no restrictions. |
| UN World Population Prospects | CC BY 3.0 IGO | cleared | CC BY 3.0 IGO; commercial use and redistribution permitted with attribution. |
| UNDP Human Development Index | CC BY 3.0 IGO | cleared | CC BY 3.0 IGO; commercial use and redistribution permitted with attribution (hdr.undp.org terms, verified 2026-07-02). |
| UNESCO World Heritage | UNESCO WHC syndication terms (permission required) | excluded | UNESCO WHC syndication requires written permission (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| US CBECS | US Public Domain | cleared | Public domain; no restrictions. |
| US Census TIGER | US Public Domain | cleared | Public domain; no restrictions. |
| US EIA | US Public Domain | cleared | Public domain; no restrictions. |
| US ISO interconnection queues (current snapshots, four operators) | Per-row - each row carries its operator's licence position in the licence/commercial_status columns | under review | UMBRELLA REGISTRATION 2026-08-29 (multi-publisher table rule - four operators must never sit under one publisher's name; the wrong-but- plausible attribution trap). Converted from caiso_queue when PJM, SPP and ISO-NE joined via gpkg_244. Rights are PER ROW - CAISO under_review (terms never confirmed), PJM derived_only (site reserves all rights, no reuse grant), SPP CLEARED on Stu's ruling 2026-08-29 ("We can serve SPP if we cite them which we do") - the grant is citation-conditioned and the per-row source and licence columns carry that citation on every served row; the terms' verbatim exception ("EXCEPT when such materials will be used ... within a commercial publication ... Any commercial use ... requires written permission") is recorded here so the ruling's basis is transparent, and a written OK from SPP remains the belt-and-braces upgrade if ever challenged, ISO-NE derived_only (copyright reserved, non-personal use restricted). Verbatim terms excerpts captured per operator in ingest/raw/iso_queues/. The umbrella's own tier is the STRICTEST of its members; serving resolves per row. PJM's api-subscription-key is PJM's own public embedded key served to every site visitor, not a held credential. |
| US Large-Scale Solar PV Database | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). Attribution to USGS/LBNL carried on /sources. |
| US state parcel fabrics (per-state publishers) | per-state; Utah CC BY 4.0 (UGRC) | cleared | Cleared per state, not nationally - the US has no national cadastre and rights, not access, are the constraint - of ~30 states surveyed only eight make an affirmative grant, the rest publish a bare disclaimer that is silent on commercial use, and absent-licence is not open-licence. Utah is cleared under UGRC's stated CC BY 4.0 (determination raw/us_parcels/ut_terms_review_20260731.txt). PERSONAL DATA - unlike European cadastres, US parcel layers routinely bundle owner names, mailing addresses and sale prices into the parcel record; ingest applies a strict FIELD ALLOWLIST so owner identity never reaches the database, the WAL or a backup, and a guard refuses to run if a mapping matches an owner/mailing/sale pattern. Utah publishes no owner name; its OWN_TYPE is an ownership CATEGORY, and its address, city, ZIP and notes columns are deliberately not read. |
| US state parcel fabrics harvested by paginated ArcGIS REST (per-state publishers) | per-state; mostly absent (disclaimer-only, internal_only); Connecticut CC0 and Victoria CC BY 4.0 cleared | cleared | Cleared per state, not nationally - each jurisdiction carries its own tier in parcel_source_registry and most are internal_only because their publishers grant nothing (absent-licence is not open-licence). Connecticut is cleared under a verbatim CC0 licenseInfo; Victoria under Vicmap CC BY 4.0. PERSONAL DATA - these layers bundle owner names, mailing addresses and CAMA valuations by default; the harvester requests ONLY an allowlisted outFields set so owner identity never reaches this process, the database, the WAL or a backup, and a guard refuses to run if a mapping matches an owner/mailing/sale pattern. REGISTRATION NOTE - this source ran unregistered from 2026-08-02 to 2026-08-06 (found during tranche 2); it was invisible to freshness and rights reporting for that window. |
| US Wind Turbine Database | US Public Domain | cleared | US federal data, public domain (verified 2026-07-11). Attribution to USGS/LBNL carried on /sources. |
| USGS Earthquake Catalog | US Public Domain | cleared | Public domain; no restrictions. |
| USGS Mineral Commodity Summaries | US Public Domain | cleared | US government work; public domain. |
| USGS Mineral Resources Data System | US Public Domain | cleared | Public domain; no restrictions. |
| USGS PAD-US 4.1 - Protected Areas Database of the United States | US public domain - the artifact's own FGDC metadata states "These data are in the public domain." | cleared | DETERMINED 2026-08-29 (T142 - the protected_areas ghost table closed at 0 rows since July): PD statement verbatim inside the artifact's FGDC XML (evidence ingest/raw/padus/ incl. the ScienceBase item JSON; download verified byte-for-byte against the advertised size). Load = ogr2ogr of the PADUS4_1Combined layer, EPSG:4326, exact count match 656,986. Geography gate: 221 features outside the CONUS+AK+HI envelope FLAGGED not rejected - Pacific marine monuments and territories legitimately sit there. |
| V-Dem & V-Party | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| VIIRS Night-time Lights | US Public Domain | cleared | Public domain; no restrictions. |
| WALS (World Atlas of Language Structures) | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Water asset layer (OpenInfraMap + GDW) | per row (ODbL / CC BY 4.0) | cleared | Constituents cleared; per-row source discloses which licence applies. |
| WHO data | WHO terms | under review | WHO data terms restrict commercial redistribution; evaluating derived-only posture. |
| World Bank Carbon Pricing Dashboard | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| World Bank energydata.info - national transmission networks | CC BY 4.0 - attribution "World Bank / energydata.info" | cleared | CLEARED 2026-08-12 (determination raw/lead_determinations/batch_determinations_20260812.txt). CC BY 4.0 stated as license_title on each CKAN record. FOUR JURISDICTIONS THAT HELD NOTHING - Laos, Montenegro, Uzbekistan and Yemen had no grid data of any kind before this, and they arrive under ONE loader because they are one publisher in one shape. PUBLISHER DEFECT GUARDED, NOT WORKED AROUND QUIETLY - the World Bank file server appends the literal string System.IO.MemoryStream after the closing brace of its GeoJSON, a .NET serialisation bug in the published artifact, affecting 6 of the 8 files; a strict parser rejects the whole document, which is why the first attempt reported three of the four countries unparseable. The trailing token is stripped because it is not part of the JSON at all, and the guard REFUSES anything else rather than truncating on a guess. VINTAGES DIFFER AND RIDE ON THE ROW - Laos 2005, Montenegro 2014, Yemen 2017, Uzbekistan undated; a 2005 network served without its date would be a claim about today. PLACE GATE TURNED FROM REJECT TO FLAG - 21 Uzbek features sit up to 3.6 degrees beyond the border in the Kyrgyz and Kazakh corridor, which Central Asia's unified power system makes plausible as real regional interconnection and equally possible as bad coordinates; they carry outside_national_envelope=true so the anomaly stays visible instead of being swallowed by a wider box or discarded. |
| World Bank PPI Database | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| World Bank Projects & Operations | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| World Bank WDI | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| World Bank WGI | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| World Database on Protected Areas | UNEP-WCMC Protected Planet terms (non-commercial, no redistribution) | excluded | Protected Planet terms prohibit commercial use and redistribution without UNEP-WCMC permission (verified 2026-07-02). Not exposed on any commercial surface; internal evaluation only. |
| World Port Index (NGA) | US Public Domain | cleared | Public domain; no restrictions. |
| WorldPop global population (R2025A 1km, constrained, UN-adjusted) | CC BY 4.0 - "WorldPop datasets are licensed under the Creative Commons Attribution 4.0 International License" (captured verbatim, ingest/raw/worldpop/) | cleared | DETERMINED 2026-08-30 (T142). CC BY 4.0 captured verbatim from the publisher's own licence file. THREE CONSTRAINTS CARRIED ON THE DATA, not just here - (1) ZONE GRAIN ONLY, population rasters clip coastal cells and understate country totals by up to 17 percent (measured on Norway), so WDI counts remain the country-grain answer and these values must never be summed to one; (2) every year 2015-2030 in R2025A is a MODEL OUTPUT interpolated and projected from two census timepoints, so 2025 is an estimate not an observation; (3) the publisher labels R2025A an alpha release that may change. All three ride in worldpop_2025_basis on every row. CROSS-VALIDATED against the independent GHSL column already in the vehicle - correlation 0.9999 across 359 zones, total ratio 1.002, which is mutual confirmation of two separately-built products rather than a single unchecked source. |
| WRI Aqueduct | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| WRI Global Power Plant Database | CC BY 4.0 | cleared | CC BY 4.0; commercial use and redistribution permitted with attribution. Attribution carried on /sources, /licensing and in response provenance. |
| Yahoo Finance (yfinance) | Yahoo terms (redistribution not licensed) | under review | Yahoo ToS do not licence redistribution; personal-use framing (verified 2026-07-11). Keyed-only, never anonymous; a licensed market-data feed is the clean production path. |
Attribution
Kortex is built on the work of the open-data community. Sources licensed CC BY, CC BY-SA, or ODbL are credited on every surface where their data appears: the source register, per-entity provenance via /v2/provenance, and the provenance.sources block of API responses. OpenStreetMap-derived layers (via OpenInfraMap) are © OpenStreetMap contributors, under ODbL.
Questions
If you are a data publisher and want to discuss how your data is represented here, or a customer with a licensing question your procurement team needs answered, write to kortex@orkora.com. Corrections are applied to the register directly and propagate to every surface within the hour.
Want to check the rights position against your own queries? A free evaluation API key is issued instantly: 14 days, faceted and insight scope, and every metered answer arrives with its sources and licences disclosed in the envelope.