{
  "schema": "dcd-open-data-dictionary/1",
  "title": "U.S. data center facility register (free download)",
  "publisher": "DataCenterData",
  "version": "2026-10-18+ed6d7733f594",
  "generated_at": "2026-10-18",
  "generated_at_basis": "Newest input date of the facility register (licensed edition, operator captures, update events and corrections). Not the build or publication time, and not evidence that every record was checked on this date.",
  "date_guide": "/methodology/data-guide/#download-dates",
  "license": {
    "name": "CC BY 4.0",
    "uri": "https://creativecommons.org/licenses/by/4.0/",
    "attribution": "DataCenterData, U.S. data center facility register (https://datacenterdata.com/data/): DataCenterData fields under CC BY 4.0",
    "suggested_citation": "DataCenterData. U.S. data center facility register, free download (data as of 2026-10-18). https://datacenterdata.com/data/. DataCenterData fields CC BY 4.0; Compute Atlas fields CC BY 4.0 (Compute Atlas by Edward Kubiak).",
    "scope": "CC BY 4.0, credit DataCenterData: the DataCenterData fields (dcd_id, public_id, site_class, site_class_label, evidence_level, evidence_tier, record_unit, county_fips, cbsa_code, cbsa_name, coordinate_precision, coordinate_source, location_precision, source_families, first_listed, latest_source_date, compute_atlas_fields, field_origins, withheld_fields) and any other field whose field_origins value is dcd_original. CC BY 4.0, credit Compute Atlas by Edward Kubiak: the fields listed in each row's compute_atlas_fields. U.S. government works (no copyright): fields whose field_origins value is public_record. dcd_url, source_urls and site_class_sources are links only, not licensed values; the linked pages and documents keep their own rights. Values listed in withheld_fields are not in this file and keep their source terms: a scalar field listed there is empty, and for a list field (operators, operator_families, aliases, capacity_values, statuses) only the entries from those sources are left out, so the field can still carry other, redistributable entries.",
    "dcd_fields": [
      "dcd_id",
      "public_id",
      "site_class",
      "site_class_label",
      "evidence_level",
      "evidence_tier",
      "record_unit",
      "county_fips",
      "cbsa_code",
      "cbsa_name",
      "coordinate_precision",
      "coordinate_source",
      "location_precision",
      "source_families",
      "first_listed",
      "latest_source_date",
      "compute_atlas_fields",
      "field_origins",
      "withheld_fields"
    ],
    "decision": "Owner decision 2026-10-28: DataCenterData-original facility fields are licensed CC BY 4.0; fields from other sources keep their own terms as stated in scope.",
    "excluded": [
      "Every value the facility register marks reference only: the operator-listed layer (names, addresses, stated operators, statuses and figures captured from operators’ pages) and DataCenterData reviewed-record fields that have not been cleared field by field. Each row’s withheld-field list names the fields they belong to (in a list such as operators, only those entries are left out); the facility page still shows them with their terms. One exception is visible: dcd_url is the address of the facility’s public DataCenterData page, built from its published page title, so for a row whose name is withheld that address can still contain a form of the name. It is a link, not a licensed value.",
      "Verbatim quotations from operator pages and other sources (held in capture records; never republished).",
      "Linked permits, filings, operator pages and other documents (linked, not republished; they keep their own rights).",
      "Selected factual fields cited to state and local agency records (shown on facility pages with their citation; not relicensed).",
      "From a Compute Atlas row that cites OpenStreetMap: its coordinate, street address and postal code. Other Compute Atlas values (for example names, cities and capacity text) pass through under Compute Atlas’s CC BY 4.0 terms; DataCenterData has not traced whether Compute Atlas drew any of them from OpenStreetMap or its other inputs. No value from DataCenterData’s own private OpenStreetMap or PNNL/IM3 lead lists is included.",
      "Coordinates the site itself does not publish (below street precision or withheld after review).",
      "Contact routes, private captures and internal review notes."
    ],
    "third_party_notices": [
      {
        "origin": "compute_atlas",
        "attribution": "Compute-infrastructure data from Compute Atlas by Edward Kubiak",
        "license": "CC BY 4.0",
        "license_uri": "https://creativecommons.org/licenses/by/4.0/",
        "source_url": "https://www.compute-atlas.com/api/facilities",
        "changes_made": "Filtered to data-center records in the 50 U.S. states; selected fields normalized for search and display; DataCenterData directory IDs added. In this download: values organized by DataCenterData facility id; coordinates, street addresses and postal codes from rows that cite OpenStreetMap removed.",
        "upstream_pass_through": "Compute Atlas reports that some of its locations derive from OpenStreetMap (ODbL) and Epoch AI (CC BY 4.0). This download removes the coordinate, street address and postal code of every Compute Atlas row that cites OpenStreetMap; any Epoch AI-derived value that remains is credited to Epoch AI under CC BY 4.0.",
        "fields": "Fields listed in each row’s compute_atlas_fields."
      },
      {
        "origin": "public_record",
        "attribution": "U.S. EPA Facility Registry Service; Approximate points: U.S. Census Bureau Geocoder (address-range interpolation, benchmark Public_AR_Current), a U.S. government work; DataCenterData review of the published street address.",
        "license": "U.S. government work, not subject to copyright (17 U.S.C. 105)",
        "license_uri": null,
        "changes_made": "Selected fields joined to DataCenterData facility ids; EPA does not endorse this dataset.",
        "upstream_pass_through": null,
        "fields": "Fields whose field_origins value is public_record."
      }
    ]
  },
  "publication": {
    "license": "CC BY 4.0",
    "license_uri": "https://creativecommons.org/licenses/by/4.0/",
    "reuse_scope": "CC BY 4.0, credit DataCenterData: the DataCenterData fields (dcd_id, public_id, site_class, site_class_label, evidence_level, evidence_tier, record_unit, county_fips, cbsa_code, cbsa_name, coordinate_precision, coordinate_source, location_precision, source_families, first_listed, latest_source_date, compute_atlas_fields, field_origins, withheld_fields) and any other field whose field_origins value is dcd_original. CC BY 4.0, credit Compute Atlas by Edward Kubiak: the fields listed in each row's compute_atlas_fields. U.S. government works (no copyright): fields whose field_origins value is public_record. dcd_url, source_urls and site_class_sources are links only, not licensed values; the linked pages and documents keep their own rights. Values listed in withheld_fields are not in this file and keep their source terms: a scalar field listed there is empty, and for a list field (operators, operator_families, aliases, capacity_values, statuses) only the entries from those sources are left out, so the field can still carry other, redistributable entries.",
    "rights_manifest": "/data/rights-manifest.json"
  },
  "record_count": 2650,
  "counts": {
    "facilities": 2650,
    "by_record_unit": {
      "building": 284,
      "campus": 170,
      "leased_presence": 47,
      "site_unknown": 2149
    },
    "by_site_class": {
      "none": 2650,
      "crypto_mining": 0,
      "ai_compute_campus": 0
    },
    "with_name": 1816,
    "with_street": 684,
    "with_county_fips": 1504,
    "by_evidence_level": {
      "dcd_record_not_in_verified_layer": 31,
      "dcd_reviewed": 155,
      "federal_record_only": 127,
      "multiple_source_families": 460,
      "operator_listing_only": 543,
      "state_local_record_only": 65,
      "third_party_listing_only": 1269
    },
    "with_coordinates": 559,
    "with_dcd_lifecycle_status": 0,
    "with_capacity_values": 1087,
    "capacity_values": 1497,
    "with_compute_atlas_values": 1728,
    "with_withheld_fields": 1389,
    "field_rights": {
      "name": {
        "dcd_original": 0,
        "compute_atlas": 1689,
        "public_record": 127,
        "withheld_source_terms": 834,
        "withheld_osm_cited_source": 0
      },
      "operators": {
        "dcd_original": 0,
        "compute_atlas": 1728,
        "public_record": 61,
        "withheld_source_terms": 1030,
        "withheld_osm_cited_source": 0
      },
      "operator_families": {
        "dcd_original": 0,
        "compute_atlas": 1728,
        "public_record": 61,
        "withheld_source_terms": 663,
        "withheld_osm_cited_source": 0
      },
      "operator_of_record": {
        "dcd_original": 0,
        "compute_atlas": 0,
        "public_record": 0,
        "withheld_source_terms": 39,
        "withheld_osm_cited_source": 0
      },
      "aliases": {
        "dcd_original": 0,
        "compute_atlas": 33,
        "public_record": 60,
        "withheld_source_terms": 516,
        "withheld_osm_cited_source": 0
      },
      "street": {
        "dcd_original": 0,
        "compute_atlas": 497,
        "public_record": 187,
        "withheld_source_terms": 470,
        "withheld_osm_cited_source": 256
      },
      "city": {
        "dcd_original": 0,
        "compute_atlas": 1525,
        "public_record": 187,
        "withheld_source_terms": 840,
        "withheld_osm_cited_source": 0
      },
      "county": {
        "dcd_original": 0,
        "compute_atlas": 1687,
        "public_record": 0,
        "withheld_source_terms": 120,
        "withheld_osm_cited_source": 0
      },
      "state": {
        "dcd_original": 0,
        "compute_atlas": 1596,
        "public_record": 187,
        "withheld_source_terms": 867,
        "withheld_osm_cited_source": 0
      },
      "postal_code": {
        "dcd_original": 0,
        "compute_atlas": 354,
        "public_record": 191,
        "withheld_source_terms": 342,
        "withheld_osm_cited_source": 241
      },
      "latitude": {
        "dcd_original": 0,
        "compute_atlas": 142,
        "public_record": 417,
        "withheld_source_terms": 51,
        "withheld_osm_cited_source": 0
      },
      "dcd_lifecycle_status": {
        "dcd_original": 0,
        "compute_atlas": 0,
        "public_record": 0,
        "withheld_source_terms": 213,
        "withheld_osm_cited_source": 0
      },
      "provider_reported_status": {
        "dcd_original": 0,
        "compute_atlas": 1728,
        "public_record": 0,
        "withheld_source_terms": 0,
        "withheld_osm_cited_source": 0
      },
      "operator_stated_status": {
        "dcd_original": 0,
        "compute_atlas": 0,
        "public_record": 0,
        "withheld_source_terms": 34,
        "withheld_osm_cited_source": 0
      },
      "capacity_values": {
        "dcd_original": 0,
        "compute_atlas": 1087,
        "public_record": 0,
        "withheld_source_terms": 561,
        "withheld_osm_cited_source": 0
      }
    }
  },
  "files": {
    "datacenters.json": {
      "sha256": "b2c3ef4505fe218b0e274add3903d83a05786b30c70b75dc3c6ef0ce536f6a8a",
      "bytes": 5440966,
      "url": "/data/datacenters.json",
      "media_type": "application/json"
    },
    "datacenters.csv": {
      "sha256": "c1581fa58cdbd159895b961af984fcc33b38c629c82abf91ebd6078a88acdb38",
      "bytes": 2362847,
      "url": "/data/datacenters.csv",
      "media_type": "text/csv; charset=utf-8"
    },
    "datacenters.geojson": {
      "sha256": "69dac7306ad6debc1a076da218a21980c1faa94c4664ae34fe6e6d9f5a930631",
      "bytes": 5560416,
      "url": "/data/datacenters.geojson",
      "media_type": "application/geo+json"
    }
  },
  "versioning": "Stable latest URLs only. Each site release replaces these files; earlier versions are not kept online. The version string is the data date plus a digest of the exported rows and is repeated in every file; cite it with the access date.",
  "format_notes": [
    "All three files carry the same facilities in the same order (dcd_id ascending); CSV is UTF-8, RFC 4180 with CRLF line ends and one header row.",
    "In CSV, list fields are joined with \"; \" and capacity_values is a JSON array. CSV text cells that begin with =, +, -, @ or a control character are prefixed with an apostrophe for spreadsheet safety; JSON and GeoJSON keep the original text.",
    "Null or empty means not published in this download, not zero. A field withheld for its source terms is listed in withheld_fields and remains on the facility page.",
    "Counts are register records (campuses, buildings, suites, unresolved presences), not unique physical sites, and not a census of U.S. data centers."
  ],
  "update_cadence": "Regenerated from the facility register with every site release; generated_at moves only when register inputs change. Changes are listed at /changes/ and in /changes.xml.",
  "fields": [
    {
      "name": "dcd_id",
      "type": "string",
      "rights": "DataCenterData identifier",
      "description": "Permanent DataCenterData facility id (dcd-fac-NNNNNN). Stable across releases; a merged or split facility keeps a redirect."
    },
    {
      "name": "public_id",
      "type": "string",
      "rights": "DataCenterData identifier",
      "description": "Short public id used in page addresses (fac-NNNNNN)."
    },
    {
      "name": "name",
      "type": "string|null",
      "rights": "current_name",
      "description": "Facility name as the best-ranked member source states it. Null when that source’s terms do not allow redistribution (see withheld_fields and the facility page)."
    },
    {
      "name": "site_class",
      "type": "array",
      "rights": "DataCenterData classification",
      "description": "crypto_mining and/or ai_compute_campus when a published, independently reviewed assertion bound to a cited official or operator source states that use (crypto-mining and AI compute campuses are included and labelled so they can be kept or dropped). Empty = unclassified: no site-class assertion is published, which does not establish that the site is, or is not, a conventional data center. Never inferred from a company name, size, tax-list membership or a provider AI label. JSON array; joined with \"; \" in CSV."
    },
    {
      "name": "site_class_label",
      "type": "string",
      "rights": "DataCenterData classification",
      "description": "Plain-language label for site_class; \"Unclassified\" when site_class is empty."
    },
    {
      "name": "site_class_sources",
      "type": "array",
      "rights": "links only",
      "description": "Source addresses of the assertions behind site_class (JSON: site_class_evidence with assertion id, source URL and capture id). The quotation stays in the capture record."
    },
    {
      "name": "site_class_evidence",
      "type": "array",
      "rights": "DataCenterData metadata",
      "description": "JSON and GeoJSON only: for each site_class value, the assertion id, source URL (a link) and capture id. Source titles and quotations are not included."
    },
    {
      "name": "evidence_level",
      "type": "string",
      "rights": "DataCenterData classification",
      "description": "Plain name for the register evidence_tier, so a discovery listing is not mistaken for a reviewed record. dcd_reviewed (tier reviewed): a member DataCenterData reviewed record is in the verified inventory layer, i.e. it cites a government, GIS, permit or docket record with locator and retrieval date; not a site inspection and not proof of current operation. multiple_source_families (tier corroborated): rows from two or more source families are joined to the facility; the register does not test whether those sources are independent of each other, so this is not corroboration. federal_record_only / state_local_record_only: one U.S. EPA FRS or one state or local government record. operator_listing_only: only the operator’s own site list. third_party_listing_only: only a Compute Atlas discovery listing (a lead, not a reviewed site). dcd_record_not_in_verified_layer (tier single_source_registry): a DataCenterData record whose inventory layer is attributed or unresolved."
    },
    {
      "name": "record_unit",
      "type": "string",
      "rights": "DataCenterData classification",
      "description": "campus, building, suite, leased_presence (an unresolved provider presence, not a separate site) or site_unknown. A campus and its buildings can each have a row; rows are register records, not unique physical sites."
    },
    {
      "name": "operators",
      "type": "array",
      "rights": "presences[]",
      "description": "Operator or company names as stated by each redistributable source, with role, stating source family and date (JSON); names joined with \"; \" in CSV. A statement is not ownership."
    },
    {
      "name": "operator_of_record",
      "type": "string|null",
      "rights": "operator_of_record",
      "description": "Operator named by a DataCenterData reviewed record, when one exists."
    },
    {
      "name": "operator_families",
      "type": "array",
      "rights": "presences[]",
      "description": "DataCenterData operator-family labels, exported only for families named by a redistributable presence or operator of record of this facility (the label names the operator, so it follows that value’s terms; field_origins gives its origin). A family known only from reference-only rows is withheld (withheld_fields operator_families:source_terms)."
    },
    {
      "name": "aliases",
      "type": "array",
      "rights": "aliases[]",
      "description": "Other names from redistributable sources (JSON: name, type, origin; CSV: names joined with \"; \")."
    },
    {
      "name": "street",
      "type": "string|null",
      "rights": "location.street",
      "description": "Street address as published by the member source shown on the facility page."
    },
    {
      "name": "city",
      "type": "string|null",
      "rights": "location.city",
      "description": "City as published by the member source."
    },
    {
      "name": "county",
      "type": "string|null",
      "rights": "location.county",
      "description": "County as published by the member source."
    },
    {
      "name": "county_fips",
      "type": "string|null",
      "rights": "DataCenterData derivation",
      "description": "5-digit Census county FIPS code, derived only from the exported county and state matched to the pinned 2024 Census county and CBSA delineation files. Blank when the county is not exported, is ambiguous, or lies outside every metro or micro area in that file. Never guessed from coordinates or city names."
    },
    {
      "name": "state",
      "type": "string|null",
      "rights": "location.state",
      "description": "Two-letter U.S. state, District of Columbia or territory code."
    },
    {
      "name": "postal_code",
      "type": "string|null",
      "rights": "location.postal_code",
      "description": "ZIP code as published by the member source."
    },
    {
      "name": "cbsa_code",
      "type": "string|null",
      "rights": "location.cbsa",
      "description": "Census core-based statistical area (metro or micro area) code assigned by DataCenterData from the county or point."
    },
    {
      "name": "cbsa_name",
      "type": "string|null",
      "rights": "location.cbsa_name",
      "description": "Census CBSA title."
    },
    {
      "name": "latitude",
      "type": "number|null",
      "rights": "location.lat",
      "description": "WGS84 latitude; present only where the DataCenterData facility page itself publishes a coordinate (exact, rooftop, parcel or street precision): a Compute Atlas provider coordinate, a U.S. government record point, or a U.S. Census Bureau Geocoder answer for a published street (a public-record point even when the street itself is withheld). Never a point copied from an operator’s own map or page, and never a state or local government address point, which the site cites from its publisher without relicensing it."
    },
    {
      "name": "longitude",
      "type": "number|null",
      "rights": "location.lon",
      "description": "WGS84 longitude; same rule as latitude."
    },
    {
      "name": "coordinate_precision",
      "type": "string|null",
      "rights": "DataCenterData metadata",
      "description": "Precision of the published coordinate (exact, rooftop, parcel, street)."
    },
    {
      "name": "coordinate_source",
      "type": "string|null",
      "rights": "DataCenterData metadata",
      "description": "compute_atlas (provider coordinate), census_geocoder (U.S. Census Bureau Geocoder address-range point reviewed by DataCenterData) or dcd_reviewed."
    },
    {
      "name": "location_precision",
      "type": "string|null",
      "rights": "DataCenterData metadata",
      "description": "Best location precision for the facility (coordinate or address), including city- or county-level records."
    },
    {
      "name": "dcd_lifecycle_status",
      "type": "string|null",
      "rights": "lifecycle_status",
      "description": "Lifecycle status from a DataCenterData reviewed record. Null when no reviewed record exists; other layers are listed separately and never reconciled into it."
    },
    {
      "name": "provider_reported_status",
      "type": "string|null",
      "rights": "status_by_layer[]",
      "description": "Compute Atlas status as reported by the provider (not independently verified). JSON statuses[] carries each layer with its date."
    },
    {
      "name": "operator_stated_status",
      "type": "string|null",
      "rights": "status_by_layer[]",
      "description": "Status stated on the operator’s own page at capture (DataCenterData compilation; the page text is not republished)."
    },
    {
      "name": "statuses",
      "type": "array",
      "rights": "status_by_layer[]",
      "description": "JSON and GeoJSON only: every redistributable status by layer with value, date and date basis. Layers are never reconciled."
    },
    {
      "name": "capacity_values",
      "type": "array",
      "rights": "capacity[], layer_values[], facts[]",
      "description": "Stated capacity, power, area and acreage values with kind, unit, scope, timing, qualifier, layer, date and source. Different kinds, scopes and sources are never added together; planned is not current, backup generation is not critical IT, a design rating is not measured load. JSON array; JSON-encoded in CSV."
    },
    {
      "name": "source_families",
      "type": "array",
      "rights": "DataCenterData metadata",
      "description": "Source layers joined in this facility: registry (DataCenterData reviewed record), federal (U.S. EPA FRS), state_local, licensed (Compute Atlas), operator (operator-listed capture)."
    },
    {
      "name": "evidence_tier",
      "type": "string",
      "rights": "DataCenterData metadata",
      "description": "The register evidence tier as computed by the register: reviewed when a member reviewed record is in the verified inventory layer, else corroborated when two or more source families are joined (independence not tested), else single_source_<family>. evidence_level gives the plain-language name."
    },
    {
      "name": "first_listed",
      "type": "string|null",
      "rights": "DataCenterData metadata",
      "description": "Date DataCenterData first allocated this facility id. Not an opening, construction or announcement date."
    },
    {
      "name": "latest_source_date",
      "type": "string|null",
      "rights": "DataCenterData metadata",
      "description": "Newest evidence date among the facility’s sources (capture, retrieval or provider update). Not proof the facility operates on that date."
    },
    {
      "name": "dcd_url",
      "type": "string",
      "rights": "link only (not a licensed value)",
      "description": "Canonical DataCenterData page with every source, unknown and rights entry. The address is built from the page title, so it can contain a form of the facility name even when the name field is withheld; it is a link, not a licensed value."
    },
    {
      "name": "source_urls",
      "type": "array",
      "rights": "links only",
      "description": "Up to 25 cited source addresses, as links. The documents themselves are not part of this dataset and keep their own rights."
    },
    {
      "name": "compute_atlas_fields",
      "type": "array",
      "rights": "DataCenterData metadata",
      "description": "Exported fields whose value comes from Compute Atlas (CC BY 4.0, attribution required)."
    },
    {
      "name": "field_origins",
      "type": "object",
      "rights": "DataCenterData metadata",
      "description": "JSON and GeoJSON only: origin of each exported content field (dcd_original, compute_atlas or public_record)."
    },
    {
      "name": "withheld_fields",
      "type": "array",
      "rights": "DataCenterData metadata",
      "description": "Fields with a value on the facility page that is left out of this download, as field:reason, where reason is source_terms (the register marks the value reference only, for example an operator-listed capture, a state or local agency record or an uncleared reviewed-record field) or osm_cited_source. A scalar field listed here is empty in this row. For a list field (operators, operator_families, aliases, capacity_values, statuses) only the entries from those sources are left out; the row can still carry other entries from redistributable sources. Omission does not mean unknown or zero."
    },
    {
      "name": "attribution",
      "type": "string",
      "rights": "notice",
      "description": "Credit lines that apply to the row: DataCenterData for its DataCenterData fields, plus Compute Atlas for the fields in compute_atlas_fields."
    }
  ],
  "csv_table_schema": {
    "fields": [
      {
        "name": "dcd_id",
        "type": "string",
        "description": "Permanent DataCenterData facility id (dcd-fac-NNNNNN). Stable across releases; a merged or split facility keeps a redirect."
      },
      {
        "name": "public_id",
        "type": "string",
        "description": "Short public id used in page addresses (fac-NNNNNN)."
      },
      {
        "name": "name",
        "type": "string",
        "description": "Facility name as the best-ranked member source states it. Null when that source’s terms do not allow redistribution (see withheld_fields and the facility page)."
      },
      {
        "name": "site_class",
        "type": "string",
        "description": "crypto_mining and/or ai_compute_campus when a published, independently reviewed assertion bound to a cited official or operator source states that use (crypto-mining and AI compute campuses are included and labelled so they can be kept or dropped). Empty = unclassified: no site-class assertion is published, which does not establish that the site is, or is not, a conventional data center. Never inferred from a company name, size, tax-list membership or a provider AI label. JSON array; joined with \"; \" in CSV."
      },
      {
        "name": "site_class_label",
        "type": "string",
        "description": "Plain-language label for site_class; \"Unclassified\" when site_class is empty."
      },
      {
        "name": "site_class_sources",
        "type": "string",
        "description": "Source addresses of the assertions behind site_class (JSON: site_class_evidence with assertion id, source URL and capture id). The quotation stays in the capture record."
      },
      {
        "name": "evidence_level",
        "type": "string",
        "description": "Plain name for the register evidence_tier, so a discovery listing is not mistaken for a reviewed record. dcd_reviewed (tier reviewed): a member DataCenterData reviewed record is in the verified inventory layer, i.e. it cites a government, GIS, permit or docket record with locator and retrieval date; not a site inspection and not proof of current operation. multiple_source_families (tier corroborated): rows from two or more source families are joined to the facility; the register does not test whether those sources are independent of each other, so this is not corroboration. federal_record_only / state_local_record_only: one U.S. EPA FRS or one state or local government record. operator_listing_only: only the operator’s own site list. third_party_listing_only: only a Compute Atlas discovery listing (a lead, not a reviewed site). dcd_record_not_in_verified_layer (tier single_source_registry): a DataCenterData record whose inventory layer is attributed or unresolved."
      },
      {
        "name": "record_unit",
        "type": "string",
        "description": "campus, building, suite, leased_presence (an unresolved provider presence, not a separate site) or site_unknown. A campus and its buildings can each have a row; rows are register records, not unique physical sites."
      },
      {
        "name": "operators",
        "type": "string",
        "description": "Operator or company names as stated by each redistributable source, with role, stating source family and date (JSON); names joined with \"; \" in CSV. A statement is not ownership."
      },
      {
        "name": "operator_of_record",
        "type": "string",
        "description": "Operator named by a DataCenterData reviewed record, when one exists."
      },
      {
        "name": "operator_families",
        "type": "string",
        "description": "DataCenterData operator-family labels, exported only for families named by a redistributable presence or operator of record of this facility (the label names the operator, so it follows that value’s terms; field_origins gives its origin). A family known only from reference-only rows is withheld (withheld_fields operator_families:source_terms)."
      },
      {
        "name": "aliases",
        "type": "string",
        "description": "Other names from redistributable sources (JSON: name, type, origin; CSV: names joined with \"; \")."
      },
      {
        "name": "street",
        "type": "string",
        "description": "Street address as published by the member source shown on the facility page."
      },
      {
        "name": "city",
        "type": "string",
        "description": "City as published by the member source."
      },
      {
        "name": "county",
        "type": "string",
        "description": "County as published by the member source."
      },
      {
        "name": "county_fips",
        "type": "string",
        "description": "5-digit Census county FIPS code, derived only from the exported county and state matched to the pinned 2024 Census county and CBSA delineation files. Blank when the county is not exported, is ambiguous, or lies outside every metro or micro area in that file. Never guessed from coordinates or city names."
      },
      {
        "name": "state",
        "type": "string",
        "description": "Two-letter U.S. state, District of Columbia or territory code."
      },
      {
        "name": "postal_code",
        "type": "string",
        "description": "ZIP code as published by the member source."
      },
      {
        "name": "cbsa_code",
        "type": "string",
        "description": "Census core-based statistical area (metro or micro area) code assigned by DataCenterData from the county or point."
      },
      {
        "name": "cbsa_name",
        "type": "string",
        "description": "Census CBSA title."
      },
      {
        "name": "latitude",
        "type": "number",
        "description": "WGS84 latitude; present only where the DataCenterData facility page itself publishes a coordinate (exact, rooftop, parcel or street precision): a Compute Atlas provider coordinate, a U.S. government record point, or a U.S. Census Bureau Geocoder answer for a published street (a public-record point even when the street itself is withheld). Never a point copied from an operator’s own map or page, and never a state or local government address point, which the site cites from its publisher without relicensing it."
      },
      {
        "name": "longitude",
        "type": "number",
        "description": "WGS84 longitude; same rule as latitude."
      },
      {
        "name": "coordinate_precision",
        "type": "string",
        "description": "Precision of the published coordinate (exact, rooftop, parcel, street)."
      },
      {
        "name": "coordinate_source",
        "type": "string",
        "description": "compute_atlas (provider coordinate), census_geocoder (U.S. Census Bureau Geocoder address-range point reviewed by DataCenterData) or dcd_reviewed."
      },
      {
        "name": "location_precision",
        "type": "string",
        "description": "Best location precision for the facility (coordinate or address), including city- or county-level records."
      },
      {
        "name": "dcd_lifecycle_status",
        "type": "string",
        "description": "Lifecycle status from a DataCenterData reviewed record. Null when no reviewed record exists; other layers are listed separately and never reconciled into it."
      },
      {
        "name": "provider_reported_status",
        "type": "string",
        "description": "Compute Atlas status as reported by the provider (not independently verified). JSON statuses[] carries each layer with its date."
      },
      {
        "name": "operator_stated_status",
        "type": "string",
        "description": "Status stated on the operator’s own page at capture (DataCenterData compilation; the page text is not republished)."
      },
      {
        "name": "capacity_values",
        "type": "string",
        "description": "Stated capacity, power, area and acreage values with kind, unit, scope, timing, qualifier, layer, date and source. Different kinds, scopes and sources are never added together; planned is not current, backup generation is not critical IT, a design rating is not measured load. JSON array; JSON-encoded in CSV."
      },
      {
        "name": "source_families",
        "type": "string",
        "description": "Source layers joined in this facility: registry (DataCenterData reviewed record), federal (U.S. EPA FRS), state_local, licensed (Compute Atlas), operator (operator-listed capture)."
      },
      {
        "name": "evidence_tier",
        "type": "string",
        "description": "The register evidence tier as computed by the register: reviewed when a member reviewed record is in the verified inventory layer, else corroborated when two or more source families are joined (independence not tested), else single_source_<family>. evidence_level gives the plain-language name."
      },
      {
        "name": "first_listed",
        "type": "date",
        "description": "Date DataCenterData first allocated this facility id. Not an opening, construction or announcement date."
      },
      {
        "name": "latest_source_date",
        "type": "date",
        "description": "Newest evidence date among the facility’s sources (capture, retrieval or provider update). Not proof the facility operates on that date."
      },
      {
        "name": "dcd_url",
        "type": "string",
        "description": "Canonical DataCenterData page with every source, unknown and rights entry. The address is built from the page title, so it can contain a form of the facility name even when the name field is withheld; it is a link, not a licensed value."
      },
      {
        "name": "source_urls",
        "type": "string",
        "description": "Up to 25 cited source addresses, as links. The documents themselves are not part of this dataset and keep their own rights."
      },
      {
        "name": "compute_atlas_fields",
        "type": "string",
        "description": "Exported fields whose value comes from Compute Atlas (CC BY 4.0, attribution required)."
      },
      {
        "name": "withheld_fields",
        "type": "string",
        "description": "Fields with a value on the facility page that is left out of this download, as field:reason, where reason is source_terms (the register marks the value reference only, for example an operator-listed capture, a state or local agency record or an uncleared reviewed-record field) or osm_cited_source. A scalar field listed here is empty in this row. For a list field (operators, operator_families, aliases, capacity_values, statuses) only the entries from those sources are left out; the row can still carry other entries from redistributable sources. Omission does not mean unknown or zero."
      },
      {
        "name": "attribution",
        "type": "string",
        "description": "Credit lines that apply to the row: DataCenterData for its DataCenterData fields, plus Compute Atlas for the fields in compute_atlas_fields."
      }
    ],
    "primaryKey": "dcd_id",
    "missingValues": [
      ""
    ]
  }
}
