{
 "columns": {
  "basis": "how the number was derived, in one sentence",
  "date": "UTC day of wapr_run: the last board built that day",
  "denominator": "for a share, what it is a share of; for a sum or count, the population (in denominator_unit) that the summed or counted items are drawn from \u2014 cranes_distinctive_sites counts sites, drawn from the sites with any filing, while its n counts filings; empty when neither applies",
  "denominator_unit": "",
  "lamp": "live / degraded / blind",
  "n": "how many items the value was computed from \u2014 or, where the value is a share of a quantity (ireland_dc_share_pct), the numerator itself in denominator_unit",
  "source": "the public record it was read from",
  "unit": "unit of value",
  "value": "the reading; empty when BLIND",
  "vintage": "the period or source state the reading DESCRIBES (a 2024 emissions year, a 2025Q4 quarter, a collection window, the register state the board recorded), which is not the day it was collected",
  "wapr_run": "the board build (generated, UTC) the row was read from"
 },
 "generated": "2026-11-02T03:25:07Z",
 "plain": "16 readings, each as a daily file. One row per UTC day a board was built \u2014 including days when nothing moved and days when the instrument was blind \u2014 so a gap is a gap and a flat line is a flat line. Value, unit, count, denominator, vintage and lamp travel together; a number never leaves this site without them. A past row the rules later derive differently is never overwritten silently: it is appended to the restatements file with the row that replaced it and why. A past row that a collector's own replay of the same collection later contradicted stays on the tape as published \u2014 it is what the board served that day \u2014 and is listed, per series, under superseded_by_source_ledger with the collector ledger line that supersedes it, the value published and the value the replay derives.",
 "restatements": {
  "file": "/series/restatements.jsonl",
  "plain": "each entry names when the restatement was made (restated_on, UTC), board_run \u2014 the board-data.json generated stamp in the working tree when the restatement was written (the board it was checked against; not necessarily a deployed board) \u2014 the series, the day, the row it superseded, the row that replaced it and the reason",
  "rows": 116
 },
 "series": {
  "announced_mw": {
   "cadence": "event-driven (hand-verified register edits)",
   "csv": "/series/announced_mw.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "degraded",
     "degraded"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 13139.25,
    "prior_vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 13139.25,
    "value_changed": false,
    "vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "prime_register",
   "last": "2026-11-02",
   "latest": {
    "basis": "sum of MW stated in a permit, docket, securities filing or exchange report; company claims excluded; 9 register rows summed, of 12 primary-verified",
    "denominator": 18,
    "denominator_unit": "register rows",
    "lamp": "degraded",
    "n": 9,
    "unit": "MW",
    "value": 13139.25,
    "vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)"
   },
   "plain": "On-site power for data centres, MW stated in a primary document",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "WAPR's own record, not a publisher's release",
    "model": "wapr_record",
    "next_release": null,
    "next_release_null_reason": "WAPR's own record, not a publisher's release",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series announced_mw latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 14,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "cranes_distinctive_sites": {
   "cadence": "daily FAA filings, calendar-year window",
   "csv": "/series/cranes_distinctive_sites.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "degraded",
     "degraded"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 350,
    "prior_vintage": "2026",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 350,
    "value_changed": false,
    "vintage": "2026",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "faa_part77",
   "last": "2026-11-02",
   "latest": {
    "basis": "\u22653 filings within 1 km AND Poisson tail p<0.01 against the 25 km background; a site is the register objects sharing one name, operator and country, each within 3 km of another member (single linkage; unnamed objects are never merged); a filing counts for every site it is within 1000 m of",
    "denominator": 680,
    "denominator_unit": "sites with any filing",
    "lamp": "degraded",
    "n": 3189,
    "unit": "sites with distinctive crane activity",
    "value": 350,
    "vintage": "2026"
   },
   "plain": "Sites with statistically distinctive crane activity this year",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for faa_part77 yet; the per-region 'Data as of' trailer every run already records (source_as_of), with every region's latest fetch state ok",
    "model": "periodic",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for faa_part77 yet; the per-region 'Data as of' trailer every run already records (source_as_of), with every region's latest fetch state ok",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series cranes_distinctive_sites latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 1,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "dmr_exceedances_12mo": {
   "cadence": "monthly filings, trailing 12 months; collected weekly",
   "csv": "/series/dmr_exceedances_12mo.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 4,
    "prior_vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 4,
    "value_changed": false,
    "vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)",
    "vintage_changed": false
   },
   "first": "2026-09-15",
   "instrument": "echo_dmr",
   "last": "2026-11-02",
   "latest": {
    "basis": "EPA's own exceedance flag on filed DMR values across all parameters",
    "denominator": 81,
    "denominator_unit": "data-centre-coded NPDES permits",
    "lamp": "live",
    "n": 16,
    "unit": "filed values over their permit limit, 12 months",
    "value": 4,
    "vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)"
   },
   "plain": "Filed DMR values that broke a permit limit in the last 12 months",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for echo_dmr yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' and 'Water Pollutant Loading Tool' rows' Date Data Extracted against the collection time, and their Expected Next Extract",
    "model": "live_query",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for echo_dmr yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' and 'Water Pollutant Loading Tool' rows' Date Data Extracted against the collection time, and their Expected Next Extract",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series dmr_exceedances_12mo latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 0,
   "rows": 35,
   "superseded_by_source_ledger": [
    {
     "date": "2026-09-15",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-15T14:23:09Z"
    },
    {
     "date": "2026-09-16",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-16T02:21:11Z"
    },
    {
     "date": "2026-09-17",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-17T14:26:24Z"
    },
    {
     "date": "2026-09-18",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-18T14:23:36Z"
    },
    {
     "date": "2026-09-19",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-19T14:22:58Z"
    },
    {
     "date": "2026-09-20",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-20T14:23:22Z"
    },
    {
     "date": "2026-09-21",
     "ledger": "data/dmr/restatements.jsonl",
     "ledger_line": 1,
     "published_in": "/dmr.json restatements",
     "restated_at": "2026-09-21T20:32:52Z",
     "superseded": {
      "value": {
       "published": "0",
       "replay": 4
      }
     },
     "vintage": "2025-09-14..2026-08-31 (collected 2026-09-15)",
     "wapr_run": "2026-09-21T20:17:57Z"
    }
   ],
   "superseded_rows_n": 7
  },
  "dmr_flow_mgd_12mo": {
   "cadence": "monthly filings, trailing 12 months; collected weekly",
   "csv": "/series/dmr_flow_mgd_12mo.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 1.247,
    "prior_vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 1.247,
    "value_changed": false,
    "vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)",
    "vintage_changed": false
   },
   "first": "2026-09-15",
   "instrument": "echo_dmr",
   "last": "2026-11-02",
   "latest": {
    "basis": "permittee-filed Discharge Monitoring Reports, parameter 50050 (flow in conduit) or 00056 (flow rate), effluent, MGD; the filed average where one was filed for that outfall and period, else the daily maximum; summed over external outfalls per period; a measurement, not a permit ceiling",
    "denominator": 81,
    "denominator_unit": "data-centre-coded NPDES permits",
    "lamp": "live",
    "n": 4,
    "unit": "MGD, sum of 12-month mean reported flows",
    "value": 1.247,
    "vintage": "2025-10-31..2026-10-31 (collected 2026-11-01)"
   },
   "plain": "Water actually discharged as filed in monthly DMRs, sum of 12-month means",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for echo_dmr yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' and 'Water Pollutant Loading Tool' rows' Date Data Extracted against the collection time, and their Expected Next Extract",
    "model": "live_query",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for echo_dmr yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' and 'Water Pollutant Loading Tool' rows' Date Data Extracted against the collection time, and their Expected Next Extract",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series dmr_flow_mgd_12mo latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 0,
   "rows": 35,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "facilities_attributed_pct": {
   "cadence": "continuous volunteer edits; read when the osmium extract is re-run (the vintage moves then)",
   "csv": "/series/facilities_attributed_pct.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 60,
    "prior_vintage": "2026-11-01",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 60,
    "value_changed": false,
    "vintage": "2026-11-01",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "facilities_osm",
   "last": "2026-11-02",
   "latest": {
    "basis": "share of mapped facilities whose operator is recorded",
    "denominator": 5041,
    "denominator_unit": "mapped facilities",
    "lamp": "live",
    "n": 3019,
    "unit": "% with operator",
    "value": 60,
    "vintage": "2026-11-01"
   },
   "plain": "Share of mapped facilities with a recorded operator",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for facilities_osm yet; Geofabrik's <region>-updates/state.txt for each extract, against extracts_as_of (OpenStreetMap is edited continuously, so an extract is behind it by nature)",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for facilities_osm yet; Geofabrik's <region>-updates/state.txt for each extract, against extracts_as_of (OpenStreetMap is edited continuously, so an extract is behind it by nature)",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series facilities_attributed_pct latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 7,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "facilities_mapped": {
   "cadence": "continuous volunteer edits; read when the osmium extract is re-run (the vintage moves then)",
   "csv": "/series/facilities_mapped.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 5041,
    "prior_vintage": "2026-11-01",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 5041,
    "value_changed": false,
    "vintage": "2026-11-01",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "facilities_osm",
   "last": "2026-11-02",
   "latest": {
    "basis": "objects tagged as data centres in the declared extracts; a frame, never a world total; 885 distinct operator names after the register's alias folding (board operators_basis)",
    "denominator": "",
    "denominator_unit": "",
    "lamp": "live",
    "n": 5041,
    "unit": "mapped facilities",
    "value": 5041,
    "vintage": "2026-11-01"
   },
   "plain": "Mapped facilities in the declared frame",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for facilities_osm yet; Geofabrik's <region>-updates/state.txt for each extract, against extracts_as_of (OpenStreetMap is edited continuously, so an extract is behind it by nature)",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for facilities_osm yet; Geofabrik's <region>-updates/state.txt for each extract, against extracts_as_of (OpenStreetMap is edited continuously, so an extract is behind it by nature)",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series facilities_mapped latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 14,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "hf_llm_new": {
   "cadence": "continuous; one row per UTC day, counting the collection windows fetched that day",
   "csv": "/series/hf_llm_new.csv",
   "delta_since_prior_row": {
    "change": null,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 634,
    "prior_vintage": "2026-10-31T14:15:04Z..2026-11-01T15:15:03Z (25.0 h)",
    "reading": "windows of different length (25 h, then 12 h) \u2014 the counts are not comparable",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 274,
    "value_changed": true,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)",
    "vintage_changed": true,
    "window_hours": [
     25.0,
     12.0
    ]
   },
   "first": "2026-09-14",
   "instrument": "hf_creation_feed",
   "last": "2026-11-02",
   "latest": {
    "basis": "creation events of repos whose pipeline tag is one of any-to-any, conversational, image-text-to-text, text-generation, text2text-generation, created after the previous watermark and fetched that UTC day; a repo deleted and re-created under the same id counts again, so this is not a count of distinct repos; windows differ in length (hours in vintage), so two rows' counts compare only over equal windows",
    "denominator": 1888,
    "denominator_unit": "repo creation events of any pipeline tag in the collection window",
    "fetch_lag_h": {
     "basis": "hours from a repo's createdAt to its first_seen (the collector run that first fetched it), over the LLM-tagged creation events counted; median, p90 (nearest rank) and max, to 0.1 h; rows is how many carry both stamps. The pipeline tag that puts a repo in the LLM population is read at first_seen and never re-read",
     "max": 12.0,
     "median": 7.7,
     "p90": 11.3,
     "rows": 274
    },
    "lamp": "live",
    "n": 274,
    "unit": "LLM-tagged repo creation events in the collection window",
    "value": 274,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)"
   },
   "plain": "LLM-tagged repo creation events on Hugging Face per collection window \u2014 a re-created id counts again (vintage gives the window and its hours)",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series hf_llm_new latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 7,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0,
   "window_kind": "per UTC day: the collection windows fetched that day, up to that day's last board"
  },
  "hf_quantized_pct": {
   "cadence": "continuous; one row per UTC day, counting the collection windows fetched that day",
   "csv": "/series/hf_quantized_pct.csv",
   "delta_since_prior_row": {
    "change": 4.57,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 40.69,
    "prior_vintage": "2026-10-31T14:15:04Z..2026-11-01T15:15:03Z (25.0 h)",
    "reading": "new vintage, value changed",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 45.26,
    "value_changed": true,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)",
    "vintage_changed": true,
    "window_hours": [
     25.0,
     12.0
    ]
   },
   "first": "2026-09-14",
   "instrument": "hf_creation_feed",
   "last": "2026-11-02",
   "latest": {
    "basis": "GGUF/AWQ/GPTQ/EXL2 name or tag over the window's creation events of repos with an LLM pipeline tag",
    "denominator": 274,
    "denominator_unit": "LLM-tagged repo creation events in the collection window",
    "fetch_lag_h": {
     "basis": "hours from a repo's createdAt to its first_seen (the collector run that first fetched it), over the LLM-tagged creation events counted; median, p90 (nearest rank) and max, to 0.1 h; rows is how many carry both stamps. The pipeline tag that puts a repo in the LLM population is read at first_seen and never re-read",
     "max": 12.0,
     "median": 7.7,
     "p90": 11.3,
     "rows": 274
    },
    "lamp": "live",
    "n": 124,
    "unit": "% of LLM-tagged repo creation events in the collection window",
    "value": 45.26,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)"
   },
   "plain": "Share of those repackaged as quantized",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series hf_quantized_pct latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 7,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0,
   "window_kind": "per UTC day: the collection windows fetched that day, up to that day's last board"
  },
  "hf_safety_stripped_pct": {
   "cadence": "continuous; one row per UTC day, counting the collection windows fetched that day",
   "csv": "/series/hf_safety_stripped_pct.csv",
   "delta_since_prior_row": {
    "change": -1.81,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 10.57,
    "prior_vintage": "2026-10-31T14:15:04Z..2026-11-01T15:15:03Z (25.0 h)",
    "reading": "new vintage, value changed",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 8.76,
    "value_changed": true,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)",
    "vintage_changed": true,
    "window_hours": [
     25.0,
     12.0
    ]
   },
   "first": "2026-09-14",
   "instrument": "hf_creation_feed",
   "last": "2026-11-02",
   "latest": {
    "basis": "name/tag match (abliterated, uncensored, unalign\u2026) over the window's creation events of repos with an LLM pipeline tag; keyword pass, not a behavioural test",
    "denominator": 274,
    "denominator_unit": "LLM-tagged repo creation events in the collection window",
    "fetch_lag_h": {
     "basis": "hours from a repo's createdAt to its first_seen (the collector run that first fetched it), over the LLM-tagged creation events counted; median, p90 (nearest rank) and max, to 0.1 h; rows is how many carry both stamps. The pipeline tag that puts a repo in the LLM population is read at first_seen and never re-read",
     "max": 12.0,
     "median": 7.7,
     "p90": 11.3,
     "rows": 274
    },
    "lamp": "live",
    "n": 24,
    "unit": "% of LLM-tagged repo creation events in the collection window",
    "value": 8.76,
    "vintage": "2026-11-01T15:15:03Z..2026-11-02T03:15:04Z (12.0 h)"
   },
   "plain": "Share of those flagged as safety-stripped",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "a continuous feed: the vintage is the fetch window itself, so there is no release for it to be the latest of",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series hf_safety_stripped_pct latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 7,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0,
   "window_kind": "per UTC day: the collection windows fetched that day, up to that day's last board"
  },
  "ireland_dc_share_pct": {
   "cadence": "quarterly periods (the grain of the data, not how often CSO releases it: see release)",
   "csv": "/series/ireland_dc_share_pct.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 23.646,
    "prior_vintage": "2025Q4",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 23.646,
    "value_changed": false,
    "vintage": "2025Q4",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "cso_ireland",
   "last": "2026-11-02",
   "latest": {
    "basis": "100 * data_centres / all_metered_electricity_consumption, both cells from MEC02 itself",
    "denominator": 8420.0,
    "denominator_unit": "GWh",
    "lamp": "live",
    "n": 1991.0,
    "unit": "% of metered electricity",
    "value": 23.646,
    "vintage": "2025Q4"
   },
   "plain": "Ireland: data centres' share of metered electricity, by quarter",
   "release": {
    "cadence_expectation": null,
    "check_source": {
     "manifest": "data/cso_ie/manifest.jsonl",
     "state": "derived",
     "ts": "2026-11-02T03:21:31Z"
    },
    "latest_published": true,
    "latest_published_basis": "ReadDataset/MEC02/JSON-stat/2.0 with no release code returns CSO's live dataset (PxStat API-Cube); its body (extension.archive and extension.reservation false) lists 44 quarters in increasing order ending 2025Q4, so 2025Q4 is the newest quarter CSO had published in MEC02 at the fetch",
    "latest_published_checked": "2026-11-02T03:21:31Z",
    "latest_published_null_reason": null,
    "model": "periodic",
    "next_release": null,
    "next_release_null_reason": "not listed in CSO's release calendar: no entry titled 'Metered Electricity Consumption' dated on or after 2026-11-02 among its 102 releases (2026-11-03..2028-01-04)",
    "publisher_latest": "2025Q4",
    "publisher_latest_as_of": "2026-07-07T11:00:00.000Z",
    "publisher_latest_as_of_meaning": "MEC02's own `updated` stamp: when CSO last touched the table (a revision or a new quarter), not the period it covers",
    "reading_matches_check": true,
    "reading_vintage_field": "series latest.vintage (the board's panels.cso_ie.latest.period, as the row read it)",
    "reading_vintage_key": "2025Q4",
    "vintage_examined": "2025Q4",
    "vintage_field": "cso-ireland.json coverage.latest",
    "vintage_key": "2025Q4"
   },
   "restated_rows": 1,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "night_solar_pct": {
   "cadence": "hourly EIA-930, trailing measurement window",
   "csv": "/series/night_solar_pct.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 0.022,
    "prior_vintage": "2025-10-01..2026-09-30",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 0.022,
    "value_changed": false,
    "vintage": "2025-10-01..2026-09-30",
    "vintage_changed": false
   },
   "first": "2026-09-14",
   "instrument": "gridmix",
   "last": "2026-11-02",
   "latest": {
    "basis": "EIA-930 hourly generation by fuel, local 01:00\u201305:00, site-weighted per operator; grids where sites sit, not power purchased",
    "denominator": 6,
    "denominator_unit": "operators on the grid-mix panel",
    "lamp": "live",
    "n": 6,
    "unit": "% solar, night window, mean over operators",
    "value": 0.022,
    "vintage": "2025-10-01..2026-09-30"
   },
   "plain": "Solar share of the grids under operators' sites, 01:00\u201305:00 local",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "a continuous hourly feed: WAPR's window ends at the last complete month before EIA's endPeriod, a window WAPR chooses by its own rule, not a release",
    "model": "continuous",
    "next_release": null,
    "next_release_null_reason": "a continuous hourly feed: WAPR's window ends at the last complete month before EIA's endPeriod, a window WAPR chooses by its own rule, not a release",
    "publisher_latest": "2026-11-01T06",
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "publisher_latest_source": {
     "field": "endPeriod",
     "manifest": "data/gridmix/manifest.jsonl",
     "meaning": "EIA-930's endPeriod, verbatim: the newest hour (UTC, hour-ending, 'YYYY-MM-DDTHH') EIA's route metadata stated when gridmix.py fetched it. WAPR's window ends at the last complete month before it, by WAPR's own rule",
     "ts": "2026-11-02T03:20:59Z"
    },
    "reading_matches_check": null,
    "reading_vintage_field": "series night_solar_pct latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 8,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "npdes_permits_518210": {
   "cadence": "irregular (permit actions); ECHO re-queried each run",
   "csv": "/series/npdes_permits_518210.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "live",
     "live"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 45,
    "prior_vintage": "as held by ECHO at 2026-11-01T15:23:42Z",
    "reading": "new vintage, same value",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 45,
    "value_changed": false,
    "vintage": "as held by ECHO at 2026-11-02T03:23:50Z",
    "vintage_changed": true
   },
   "first": "2026-09-15",
   "instrument": "discharge",
   "last": "2026-11-02",
   "latest": {
    "basis": "Clean Water Act permits whose holder reports the data-processing sector code; a floor, not an inventory; 0 carry a reported flow; 52 jurisdictions queried (states, DC and Puerto Rico); 0 of them failed",
    "denominator": "",
    "denominator_unit": "",
    "lamp": "live",
    "n": 45,
    "unit": "NPDES permits (NAICS 518210)",
    "value": 45,
    "vintage": "as held by ECHO at 2026-11-02T03:23:50Z"
   },
   "plain": "US discharge permits held by facilities reporting the data-centre sector code",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for discharge yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' row's Date Data Extracted against the collection time, and its Expected Next Extract",
    "model": "live_query",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for discharge yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' row's Date Data Extracted against the collection time, and its Expected Next Extract",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series npdes_permits_518210 latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 13,
   "rows": 35,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "npdes_reported_flow_mgd": {
   "cadence": "irregular (permit actions); ECHO re-queried each run",
   "csv": "/series/npdes_reported_flow_mgd.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "degraded",
     "degraded"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 0,
    "prior_vintage": "as held by ECHO at 2026-11-01T15:23:42Z",
    "reading": "new vintage, same value",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 0,
    "value_changed": false,
    "vintage": "as held by ECHO at 2026-11-02T03:23:50Z",
    "vintage_changed": true
   },
   "first": "2026-09-15",
   "instrument": "discharge",
   "last": "2026-11-02",
   "latest": {
    "basis": "sum of average flow recorded on the permits that carry one; a permit figure, not a meter reading",
    "denominator": 45,
    "denominator_unit": "permits",
    "lamp": "degraded",
    "n": 0,
    "unit": "MGD",
    "value": 0,
    "vintage": "as held by ECHO at 2026-11-02T03:23:50Z"
   },
   "plain": "Average flow recorded on those permits, million gallons per day",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no release check implemented for discharge yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' row's Date Data Extracted against the collection time, and its Expected Next Extract",
    "model": "live_query",
    "next_release": null,
    "next_release_null_reason": "no release check implemented for discharge yet; one GET of ECHO's About-the-Data page per collection (https://echo.epa.gov/resources/echo-data/about-the-data): its 'ICIS-NPDES' row's Date Data Extracted against the collection time, and its Expected Next Extract",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series npdes_reported_flow_mgd latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 7,
   "rows": 35,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "onsite_islanded_mw": {
   "cadence": null,
   "csv": "/series/onsite_islanded_mw.csv",
   "delta_since_prior_row": {
    "change": 0.0,
    "date": "2026-11-02",
    "hours_between_runs": 11.97,
    "lamps": [
     "degraded",
     "degraded"
    ],
    "prior_date": "2026-11-01",
    "prior_run": "2026-11-01T15:26:40Z",
    "prior_value": 7525.85,
    "prior_vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)",
    "reading": "no change",
    "run": "2026-11-02T03:25:07Z",
    "state": "ok",
    "value": 7525.85,
    "value_changed": false,
    "vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)",
    "vintage_changed": false
   },
   "first": "2026-09-15",
   "instrument": null,
   "last": "2026-11-02",
   "latest": {
    "basis": "MW whose own primary document states the project has NO utility connection, of 13,139.25 MW announced across 6 grid relations. ANNOUNCED, NEVER RUNNING: operating_mw is a separate reading and is BLIND until a primary document states running MW; a permit ceiling is a limit, not a reading",
    "denominator": 9,
    "denominator_unit": "on-site power rows the board counted",
    "lamp": "degraded",
    "n": 3,
    "unit": "MW announced with no utility connection",
    "value": 7525.85,
    "vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)"
   },
   "plain": "Announced on-site power a primary document says has NO utility connection \u2014 announced, never running",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "no series\u2192instrument mapping for onsite_islanded_mw in build/release.py",
    "model": null,
    "next_release": null,
    "next_release_null_reason": "no series\u2192instrument mapping for onsite_islanded_mw in build/release.py",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 8,
   "rows": 35,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "operating_mw": {
   "cadence": "event-driven (hand-verified register edits)",
   "csv": "/series/operating_mw.csv",
   "delta_since_prior_row": {
    "date": "2026-11-02",
    "prior_date": "2026-11-01",
    "prior_lamp": "blind",
    "prior_value": "",
    "reading": "latest row is BLIND \u2014 no reading to compare",
    "run": "2026-11-02T03:25:07Z",
    "state": "blind"
   },
   "first": "2026-09-14",
   "instrument": "prime_register",
   "last": "2026-11-02",
   "latest": {
    "basis": "MW with a primary OPERATING record (air-permit operating statement, CAMPD unit, EIA-923 generation); announcements and permit ceilings never count",
    "denominator": 18,
    "denominator_unit": "register rows",
    "lamp": "blind",
    "n": 0,
    "unit": "MW",
    "value": "",
    "vintage": "register as last edited 2026-09-15T07:40:16Z (the board's own record)"
   },
   "plain": "On-site power with a primary OPERATING record, MW \u2014 BLIND until one exists",
   "release": {
    "cadence_expectation": null,
    "check_source": null,
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": null,
    "latest_published_null_reason": "WAPR's own record, not a publisher's release",
    "model": "wapr_record",
    "next_release": null,
    "next_release_null_reason": "WAPR's own record, not a publisher's release",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": null,
    "reading_vintage_field": "series operating_mw latest.vintage is a formatted string, not a raw key",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": null,
    "vintage_key": null
   },
   "restated_rows": 14,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  },
  "va_nox_t": {
   "cadence": "annual emissions year",
   "csv": "/series/va_nox_t.csv",
   "delta_since_prior_row": {
    "date": "2026-11-02",
    "prior_date": "2026-11-01",
    "prior_lamp": "blind",
    "prior_value": "",
    "reading": "latest row is BLIND \u2014 no reading to compare",
    "run": "2026-11-02T03:25:07Z",
    "state": "blind"
   },
   "first": "2026-09-14",
   "instrument": "va_deq_sites",
   "last": "2026-11-02",
   "latest": {
    "basis": "BLIND \u2014 no input this run: board panel va.nox_sum_t",
    "denominator": "",
    "denominator_unit": "",
    "lamp": "blind",
    "n": "",
    "unit": "short tons NOx/yr",
    "value": "",
    "vintage": ""
   },
   "plain": "Virginia data-centre generator NOx per year, in the unit DEQ's emissions table states (each row's unit column), by vintage",
   "release": {
    "cadence_expectation": null,
    "check_source": {
     "layer": 292,
     "manifest": "data/va_deq/manifest.jsonl",
     "state": "fetch_failed",
     "ts": "2026-11-02T03:15:13Z"
    },
    "latest_published": null,
    "latest_published_basis": null,
    "latest_published_checked": "2026-11-02T03:15:13Z",
    "latest_published_null_reason": "table 292 was not read this run (the fetch failed)",
    "model": "live_query",
    "next_release": null,
    "next_release_null_reason": "DEQ publishes no release calendar for table 292; the table overwrites itself in place",
    "publisher_latest": null,
    "publisher_latest_as_of": null,
    "publisher_latest_as_of_meaning": null,
    "reading_matches_check": false,
    "reading_vintage_field": "series latest.vintage (the board's panels.va.nox_year, as the row read it)",
    "reading_vintage_key": null,
    "vintage_examined": null,
    "vintage_field": "va_deq manifest layer-292 row nox_year (the board's panels.va.nox_year)",
    "vintage_key": null
   },
   "restated_rows": 8,
   "rows": 36,
   "superseded_by_source_ledger": [],
   "superseded_rows_n": 0
  }
 },
 "source_ledgers": {
  "plain": "collector ledgers (data/<collection>/restatements.jsonl) read for tape rows their replays superseded; an unreadable line is listed, not skipped",
  "read": [
   "data/dmr/restatements.jsonl"
  ],
  "rows_naming_no_series_here": 0,
  "unreadable_lines": []
 }
}