{
 "schema_version": 2,
 "release_date": "2026-09-13",
 "status": "research-release",
 "certification": "not externally certified",
 "code_revision": "6f3d247",
 "cohort_calibration": {
  "publisher": "United Nations Department of Economic and Social Affairs, Population Division",
  "release": "World Population Prospects 2024, Online Edition",
  "source_file": "WPP2024_POP_F01_1_POPULATION_SINGLE_AGE_BOTH_SEXES.xlsx",
  "source_sha256": "78109efb998fee10bd0c501844e04cafd0922836239e00aa90d9f35fc828a94a",
  "method": "WPP mid-year single-age population of the cohort in each calendar year, divided by the model birth cohort; interpolated linearly between 1 July reference points",
  "coverage": {
   "canonical_entities": 237,
   "calibrated_entities": 236,
   "explicit_fallback_entities": [
    "Kosovo"
   ]
  },
  "reproduction_check": "Model reproduces WPP 2024's own Japan population aged 13-19 on 1 July 2024 within 0.65%."
 },
 "births_input": {
  "source": "Our World in Data export of UN World Population Prospects births by year",
  "file": "number-of-births-per-year.csv",
  "sha256": "5f9f5873b2005ed90e85447d56ff4569244073d9b059ff8bbbe299110b01f9f0",
  "provenance": "data/sources/existing-inputs-provenance.md"
 },
 "seasonality": {
  "method": "Monthly shares averaged over up to 5 most recent complete years >= 2005; within-month distribution is uniform per day; regional fallback is a birth-weighted mean of direct-data countries in the smallest M49 group available.",
  "built_utc": "2026-09-13T02:17:56.372856+00:00",
  "tiers": [
   {
    "tier": "hfd_stff",
    "label": "Human Fertility Database, Short-Term Fertility Fluctuations (monthly births)",
    "countries": 43,
    "share_of_2005_2013_births": 0.1031
   },
   {
    "tier": "eurostat_demo_fmonth",
    "label": "Eurostat demo_fmonth (live births by month)",
    "countries": 17,
    "share_of_2005_2013_births": 0.0132
   },
   {
    "tier": "undy_table55",
    "label": "UN Demographic Yearbook table 55 (live births by month of occurrence)",
    "countries": 79,
    "share_of_2005_2013_births": 0.1947
   },
   {
    "tier": "published_shares",
    "label": "Published national tables and peer-reviewed monthly shares (Dorélien 2016; UNSD)",
    "countries": 6,
    "share_of_2005_2013_births": 0.1243
   },
   {
    "tier": "regional_fallback",
    "label": "Regional fallback: birth-weighted mean of direct-data countries in the smallest available M49 group",
    "countries": 92,
    "share_of_2005_2013_births": 0.5647
   }
  ],
  "direct_data_share_of_births": 0.4353,
  "sources": {
   "eurostat_demo_fmonth": {
    "file": "data/sources/seasonality/eurostat_demo_fmonth.csv",
    "sha256": "e47001dd948d29151c994c971101dcfc3afe725a8369f49aa510e64ab5603658"
   },
   "hfd_stff": {
    "file": "data/sources/seasonality/hfd_births_by_month.csv",
    "sha256": "32570791903c9a855715042bf0c44a9ec8499db1775d44898d851326ed7effe8"
   },
   "undy_table55": {
    "file": "data/sources/seasonality/undy_births_by_month.csv",
    "sha256": "f4edec72204cc9a821dde3377e88a0305de0329ddc2001eaf78941b49ec683b1"
   }
  },
  "published_shares_provenance": "data/sources/seasonality/tier4/published.provenance.json",
  "largest_entities_on_fallback": [
   "India",
   "China",
   "Indonesia",
   "Bangladesh",
   "Viet Nam",
   "Afghanistan",
   "Iraq",
   "Sudan"
  ],
  "fallback_note": "India and China publish no openly available monthly births table (civil-registration reports, SRS, NFHS-5 and NBS census volumes were checked; see data/sources/seasonality/tier4/india_notes.md and china_notes.md). They use the Southern Asia / Eastern Asia regional profiles.",
  "fallback_entities": [
   "Afghanistan",
   "Angola",
   "Bangladesh",
   "Belize",
   "Benin",
   "Bhutan",
   "Bolivia",
   "Bonaire Sint Eustatius and Saba",
   "Burkina Faso",
   "Burundi",
   "Cambodia",
   "Cameroon",
   "Central African Republic",
   "Chad",
   "China",
   "Comoros",
   "Congo",
   "Cook Islands",
   "Cote d'Ivoire",
   "Djibouti",
   "Dominica",
   "Dominican Republic",
   "East Timor",
   "Ecuador",
   "Equatorial Guinea",
   "Eritrea",
   "Eswatini",
   "Falkland Islands",
   "Fiji",
   "Gabon",
   "Gambia",
   "Ghana",
   "Grenada",
   "Guinea",
   "Guinea-Bissau",
   "Guyana",
   "Haiti",
   "Honduras",
   "India",
   "Indonesia",
   "Iraq",
   "Jordan",
   "Kiribati",
   "Laos",
   "Lesotho",
   "Liberia",
   "Libya",
   "Madagascar",
   "Malawi",
   "Mali",
   "Marshall Islands",
   "Mauritania",
   "Micronesia (country)",
   "Mozambique",
   "Myanmar",
   "Namibia",
   "Nauru",
   "Nepal",
   "Niger",
   "North Korea",
   "Northern Mariana Islands",
   "Papua New Guinea",
   "Paraguay",
   "Peru",
   "Rwanda",
   "Saint Barthelemy",
   "Saint Kitts and Nevis",
   "Saint Lucia",
   "Samoa",
   "Saudi Arabia",
   "Senegal",
   "Sierra Leone",
   "Sint Maarten (Dutch part)",
   "Solomon Islands",
   "Somalia",
   "South Sudan",
   "Sudan",
   "Syria",
   "Tajikistan",
   "Togo",
   "Tokelau",
   "Tonga",
   "Turkmenistan",
   "Tuvalu",
   "Vanuatu",
   "Vatican",
   "Vietnam",
   "Wallis and Futuna",
   "Western Sahara",
   "Yemen",
   "Zambia",
   "Zimbabwe"
  ],
  "within_month_allocation": "uniform per day within each month"
 },
 "hour_of_birth": {
  "source": "US CDC/NCHS Data Brief No. 200 (2015), \"When Are Babies Born\", 2013 birth certificate data, Total column",
  "scope": "One common hour-of-day curve applied to every country; a small effect on any daily total. No other country publishes an open numeric hour-of-birth table.",
  "provenance": {
   "generated_utc": "2026-09-13T01:37:32Z",
   "purpose": "Provenance for a properly sourced replacement of the top-level birth-by-hour.csv (which shipped with no recorded source). Search covered US CDC/NCHS, UK ONS/NHS-linked research, and Japan e-Stat; only the US CDC table was found in openly downloadable, machine-extractable tabular form.",
   "included_in_hour_of_birth_curves_csv": [
    {
     "source_id": "us_cdc_nchs_db200_2013",
     "title": "When Are Babies Born: Morning, Noon, or Night? Birth Certificate Data for 2013 (NCHS Data Brief No. 200)",
     "authors": "T.J. Mathews, S.C. Curtin",
     "publisher": "CDC / National Center for Health Statistics (NCHS), National Vital Statistics System",
     "published": "May 2015",
     "iso3": "USA",
     "year_range": "2013",
     "url_landing": "https://www.cdc.gov/nchs/products/databriefs/db200.htm",
     "url_report_pdf": "https://www.cdc.gov/nchs/data/databriefs/db200.pdf",
     "url_data_table_pdf": "https://www.cdc.gov/nchs/data/databriefs/db200_table.pdf",
     "retrieved_utc": "2026-09-13T01:34:54Z",
     "raw_files": [
      {
       "path": "raw/cdc_nchs_db200.pdf",
       "sha256": "962e4a9996c1f184ad0051145f480a5683606578d9edfccc3ce3eaecef59a7a1",
       "bytes": 617678,
       "content": "Full data brief (narrative + Figures 1-4 as charts)"
      },
      {
       "path": "raw/cdc_nchs_db200_table.pdf",
       "sha256": "da06ed07facd96cd2bd1921af25b5212b03534eff6e7aaa6446810605f407c9c",
       "bytes": 114595,
       "content": "Data tables for Figures 1-4, machine-readable text (this is what curves were digitized from)"
      }
     ],
     "population_covered": "Live births, birth-certificate reporting area of 41 US states plus the District of Columbia that had adopted the 2003-revised birth certificate by January 2013, representing 90% of all US births in 2013. Includes ALL modes of delivery and ALL places of delivery (hospital + out-of-hospital) combined -- this is the 'Total' column of Data table for Figure 1.",
     "curve_used": "Data table for Figure 1, 'Total' column: percent distribution of births by hour of delivery (all days of week combined), 2013.",
     "other_columns_in_source_not_used": [
      "Monday-Friday vs Saturday-Sunday breakdown (Fig 1)",
      "Cesarean / induced vaginal / noninduced vaginal breakdown (Fig 2)",
      "No-trial-of-labor vs trial-of-labor cesarean breakdown (Fig 3)",
      "In-hospital vs out-of-hospital breakdown (Fig 4)"
     ],
     "notes": [
      "Hour labels in the source run 'Midnight' (00:00-00:59) through '11 p.m.' (23:00-23:59); mapped directly to hour=0..23 in the output CSV, i.e. hour N covers N:00-N:59 local clock time at place of birth.",
      "The 24 published percentages sum to 100.1 (source rounding to 1 decimal place); the 'share' column in hour_of_birth_curves.csv is the CDC percentage renormalized so the 24 values sum to exactly 1.0 (each value divided by 100.1, with a final +0.000000...1 rounding correction applied to hour 23 so the stored 6-decimal values sum to 1.0 exactly).",
      "This 'Total' column is numerically IDENTICAL (all 24 values) to the pre-existing, unsourced /home/opc/live_teen_population/birth-by-hour.csv. That file was in fact this CDC table, just shipped without any citation -- see comparison.md."
     ]
    }
   ],
   "found_but_not_included": [
    {
     "id": "uk_nhs_england_martin_2018",
     "title": "Timing of singleton births by onset of labour and mode of birth in NHS maternity units in England, 2005-2014: A study of linked birth registration, birth notification, and hospital episode data",
     "authors": "Martin P, Cortina-Borja M, Newburn M, Harper G, Gibson R, Dodwell M, Dattani N, Macfarlane A",
     "publisher": "PLOS ONE 13(6): e0198183, using linked ONS birth registration + birth notification + NHS Digital Maternity HES data",
     "doi": "10.1371/journal.pone.0198183",
     "url": "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0198183",
     "url_pdf_used": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0198183&type=printable",
     "retrieved_utc": "2026-09-13T01:35:46Z",
     "raw_files": [
      {
       "path": "raw/plos_pone0198183.pdf",
       "sha256": "8323c61db99bff1197a2ed854da7c1248150666f975ae8ea1d4b70902e094f64",
       "bytes": 4005850
      }
     ],
     "population_covered": "5,093,615 singleton births in NHS maternity units in England, 2005-2014 (excludes home/out-of-hospital births and births with incomplete onset-of-labour/mode-of-birth/gestational-age data -- about 15.9% of the linked dataset excluded for missing fields).",
     "why_not_included_in_curves_csv": "The paper reports hour-of-birth only as line charts (Fig 2 = all singleton births by hour across the week; Figs 3-15 = breakdowns by onset of labour / mode of birth), not as a numeric table of percent-by-hour. No supporting-information file (S1-S4 Appendices, all DOCX) contains a machine-readable all-births-by-hour table either -- S4 Appendix is restricted to a single weekday (non-holiday Thursday) and to the eight onset/mode subgroups, not a single 24-hour total curve. Digitizing pixel positions from the Fig 2 chart would fabricate false precision, so no hour-by-hour values were extracted from this source.",
     "usable_facts_extracted_without_digitizing": [
      "48.1% of births in the England study dataset occurred in the 12 hours from 6pm to 6am, versus 39.5% in the equivalent US 2013 CDC data (a personal communication from S. Curtin cited in the paper) -- used as a coarse day/night sanity check in comparison.md, not as a per-hour value.",
      "Spontaneous vaginal births after spontaneous labour onset (just over half of England's total) peak around 4:00 am, roughly sinusoidal, matching the pre-obstetric-intervention 'night birth' pattern seen in older studies (Birmingham 1950-51, Madrid 1887-92, New York 1929/1936).",
      "Elective and emergency caesareans, and induced-then-vaginal births, are concentrated on weekday daytime/pre-midnight hours respectively -- these obstetric-intervention effects are what shift England's overall curve away from a pure night-peaked shape, similarly to the US."
     ]
    },
    {
     "id": "us_uk_nl_comparison_martin_2023",
     "title": "The natural pattern of birth timing and gestational age in the U.S. compared to England, and the Netherlands",
     "publisher": "PLOS ONE, e0278856",
     "doi": "10.1371/journal.pone.0278856",
     "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9847908/",
     "url_pdf_used": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0278856&type=printable",
     "retrieved_utc": "2026-09-13T01:37:16Z",
     "raw_files": [
      {
       "path": "raw/plos_pone0278856.pdf",
       "sha256": "c30e64d2ada5799e66ce38512e62adfcba68841027931ddffd2daff47e30442c",
       "bytes": 1539287
      }
     ],
     "population_covered": "Home births and unmedicated (no induction/augmentation) vaginal hospital births only, compared across US (2014), Netherlands (2014), England (2008-2010) -- i.e. deliberately the least-medicalized subset of births in each country, NOT all births.",
     "why_not_included_in_curves_csv": "Hour-of-birth is again presented only as charts (Fig 3, Fig 4a-4c), no numeric table, no downloadable supplementary data table with hour-by-hour values. Also explicitly not an all-births/general-population curve by design (excludes induced and augmented labours), so it would not be an apples-to-apples replacement for the model's global curve even if digitized.",
     "usable_facts_extracted_without_digitizing": [
      "All three countries show a similar shape for this unmedicated subset: peak roughly 1:00-5:59 am, trough roughly 2:00-4:59 pm -- consistent with the spontaneous-birth pattern in the England 2018 study and with the pre-1970s US pattern, reinforcing that the CDC 2013 'Total' curve's daytime peak is specifically an artifact of modern obstetric intervention (scheduled inductions/c-sections), not a universal biological pattern."
     ]
    },
    {
     "id": "japan_estat_search",
     "title": "No open hour-of-birth dataset located",
     "notes": "Searched e-Stat (Portal Site of Official Statistics of Japan) Vital Statistics of Japan Natality tables for an hour-of-birth breakdown (出生時刻). e-Stat publishes extensive natality tables (birth order, maternal age, birth weight, etc.) but no hour-of-day tabulation was found in the results returned by e-Stat's search/browse interface as of this session. Not pursued further (would require navigating e-Stat's Japanese-language table browser interactively, which was out of scope for a self-contained data task); flagged here as an open gap rather than silently dropped.",
     "retrieved_utc": "2026-09-13T01:35:00Z"
    }
   ],
   "methodology": {
    "extraction": "The two included-source PDFs (cdc_nchs_db200.pdf, cdc_nchs_db200_table.pdf) were read directly (no OCR/digitizing tool available on this host -- pdftotext/poppler-utils and PyPDF2/pdfplumber/fitz are not installed; the Read tool's built-in PDF text extraction was used instead) and the 'Total' column of the Data table for Figure 1 was transcribed by hand-checking every value against both the table PDF and the matching chart in the main report PDF (Figure 1) before being written into hour_of_birth_curves.csv.",
    "network_access": "This host has outbound internet access via curl; all downloads used curl with no authentication, no cookies beyond default redirects.",
    "cpu_note": "All work here was I/O (download) and simple arithmetic (percent -> share normalization); no heavy compute was run, per the 'no CPU stress on opc' rule."
   }
  },
  "file_sha256": "41143a5bb82b8fa572a6b7235bec998ee13c855ca31b6aec1bc81b7aa63a6bac"
 },
 "timezone_allocation": {
  "population_source": "GHS-POP R2023A, epoch 2025, global 1 km",
  "population_raster_sha256": "db9cdc20a7ab6d383f63b10b70929887fbd96200452cd48e4b6515c1a52b0f75",
  "timezone_boundaries": "Timezone Boundary Builder 2026c timezones-now",
  "timezone_geometry_sha256": "40471fed0621974b579e72ae3cfd68a9756c6c2894c348d738c6d6bec0139545",
  "method": "population-raster zonal sum at cell centres, normalized per canonical entity",
  "canonical_entities": 237,
  "single_zone_overrides": 9,
  "validation": "all 237 canonical entities have positive weights totaling exactly 100%"
 },
 "validation": {
  "method": "The model is evaluated at each national source's own reference date and compared with that source's published population aged 13-19. None of these sources is used in calibration.",
  "countries": 51,
  "country_years": 148,
  "median_abs_pct_latest_year": 2.35,
  "within_1pct": 15,
  "within_2pct": 21,
  "within_5pct": 39,
  "headline": {
   "USA": {
    "iso3": "USA",
    "year": 2024,
    "reference_date": "2024-07-01",
    "national": 30835278.0,
    "source": "US Census Bureau Vintage 2024",
    "model": 31529448,
    "diff": 694170,
    "diff_pct": 2.25
   },
   "JPN": {
    "iso3": "JPN",
    "year": 2024,
    "reference_date": "2024-10-01",
    "national": 7580000.0,
    "source": "Statistics Bureau of Japan (thousands)",
    "model": 7879036,
    "diff": 299036,
    "diff_pct": 3.95
   },
   "GBR": {
    "iso3": "GBR",
    "year": 2024,
    "reference_date": "2024-06-30",
    "national": 5825084.0,
    "source": "ONS mid-year estimates",
    "model": 5770666,
    "diff": -54418,
    "diff_pct": -0.93
   },
   "DEU": {
    "iso3": "DEU",
    "year": 2025,
    "reference_date": "2025-01-01",
    "national": 5505138.0,
    "source": "Eurostat demo_pjan (1 January)",
    "model": 5406812,
    "diff": -98326,
    "diff_pct": -1.79
   },
   "FRA": {
    "iso3": "FRA",
    "year": 2025,
    "reference_date": "2025-01-01",
    "national": 5990735.0,
    "source": "Eurostat demo_pjan (1 January)",
    "model": 5790526,
    "diff": -200209,
    "diff_pct": -3.34
   },
   "CAN": {
    "iso3": "CAN",
    "year": 2024,
    "reference_date": "2024-07-01",
    "national": 3209005.0,
    "source": "Statistics Canada 17-10-0005-01",
    "model": 3040459,
    "diff": -168546,
    "diff_pct": -5.25
   },
   "AUS": {
    "iso3": "AUS",
    "year": 2025,
    "reference_date": "2025-06-30",
    "national": 2379337.0,
    "source": "ABS 3101.0 Table 59",
    "model": 2317728,
    "diff": -61609,
    "diff_pct": -2.59
   }
  },
  "definitional_outliers": {
   "CYP": "WPP counts the whole island; Eurostat covers the government-controlled area only.",
   "MDA": "WPP includes Transnistria; national statistics exclude it.",
   "XKX": "Kosovo has no WPP row; declared calibration fallback.",
   "ALB": "Large emigration; WPP and the national register disagree on residents.",
   "GEO": "Breakaway regions counted differently by WPP and national statistics."
  },
  "note": "WPP 2024 itself sits above national figures for several well-measured countries (Japan +3.3%, United States +1.9% for ages 13-19 in 2024). The model reproduces WPP's own Japan figure within 0.65% (tests/test_wpp_calibration.py), so most of the remaining gap to national statistics is inherited from the UN source, not introduced by this model.",
  "artifact": "model_vs_national.csv",
  "generated_utc": "2026-09-13T02:18:03.364096Z",
  "sources": [
   {
    "file": "data/validation/eurostat_demo_pjan.provenance.json",
    "record": {
     "source": "Eurostat demo_pjan (Population on 1 January by age and sex)",
     "url": "https://ec.europa.eu/eurostat/api/dissemination/statistics/1.0/data/demo_pjan?format=JSON&lang=EN&sex=T&age=Y13&age=Y14&age=Y15&age=Y16&age=Y17&age=Y18&age=Y19&sinceTimePeriod=2015",
     "retrieved_utc": "2026-09-13T01:36:23Z",
     "raw_sha256": "25b9e4103af25d56304c84ef1d05d9189e28fb250dc52d63691eb1f40b74f304",
     "raw_bytes": 60313,
     "countries": [
      "ALB",
      "AND",
      "ARM",
      "AUT",
      "AZE",
      "BEL",
      "BGR",
      "BIH",
      "BLR",
      "CHE",
      "CYP",
      "CZE",
      "DEU",
      "DNK",
      "ESP",
      "EST",
      "FIN",
      "FRA",
      "GBR",
      "GEO",
      "GRC",
      "HRV",
      "HUN",
      "IRL",
      "ISL",
      "ITA",
      "LIE",
      "LTU",
      "LUX",
      "LVA",
      "MCO",
      "MDA",
      "MKD",
      "MLT",
      "MNE",
      "NLD",
      "NOR",
      "POL",
      "PRT",
      "ROU",
      "RUS",
      "SMR",
      "SRB",
      "SVK",
      "SVN",
      "SWE",
      "TUR",
      "UKR",
      "XKX"
     ],
     "year_range": [
      2015,
      2025
     ],
     "notes": "population on 1 January, ages in completed years"
    }
   },
   {
    "file": "data/validation/northamerica.provenance.json",
    "record": {
     "usa": {
      "source": "U.S. Census Bureau, Population and Housing Unit Estimates, Vintage 2024 (national, age-sex, resident population)",
      "url": "https://www2.census.gov/programs-surveys/popest/datasets/2020-2024/national/asrh/nc-est2024-agesex-res.csv",
      "retrieved_utc": "2026-09-13T01:37:38Z",
      "raw_sha256": "fbb151e8ae8554283dcbf192e6b420658ea444ee7d8849988aa1d5615183edf1",
      "raw_bytes": 16171,
      "notes": "Full national single-year-of-age file (SEX=0 both sexes, SEX=1 male, SEX=2 female; AGE 0-84 then 85 = 85 and over). July 1 postcensal estimates for 2020-2024 (POPESTIMATE2020..POPESTIMATE2024); ESTIMATESBASE2020 column (4/1/2020 base, not a July 1 estimate) was excluded. Kept in full (16KB) since it is small; no filtering needed. Filtered to SEX=0, AGE 13-19 for usa_census_13_19.csv."
     },
     "canada": {
      "source": "Statistics Canada, Table 17-10-0005-01, Population estimates on July 1, by age and gender",
      "url": "https://www150.statcan.gc.ca/t1/wds/rest/getFullTableDownloadCSV/17100005/en (resolved to https://www150.statcan.gc.ca/n1/tbl/csv/17100005-eng.zip)",
      "retrieved_utc": "2026-09-13T01:37:38Z",
      "raw_sha256": "a534e159ce82464fc5d67a8be97dd2912cb83ce3a21a5d8c6f072a8d7668e01a",
      "raw_bytes": 3806064,
      "notes": "WDS getFullTableDownloadCSV endpoint requires the bare 8-digit product id (17100005, not 1710000501 and not /en/17100005 - path order is /{productId}/{lang}); it returns a JSON pointer to the actual zip. The full table zip (3.8MB compressed, ~38.8MB uncompressed 17100005.csv) covers all provinces/territories, all genders (Total-gender/Men+/Women+), single ages and age bands, 1971-2025. Per task instructions this raw zip is kept (raw_sha256/raw_bytes above describe the zip itself), but the 38.8MB uncompressed CSV was NOT kept on disk after filtering - instead a filtered raw extract (GEO=Canada, Gender='Total - gender', Age group in {13..19 years}, REF_DATE in 2020-2024, 35 rows) was written to raw/canada_17100005_filtered_raw.csv (sha256 recorded separately below) and canada_statcan_13_19.csv was derived from that extract. 2025 (REF_DATE=2025) data exists in the table but was excluded so both countries report the same 2020-2024 window; no explicit preliminary/STATUS flag was present on any year in this table (StatCan does not flag it in this cube's STATUS column, per 17100005_MetaData.csv). Metadata file kept at raw/17100005_MetaData.csv."
     },
     "canada_filtered_extract": {
      "source": "Filtered extract of Statistics Canada Table 17-10-0005-01 (see 'canada' entry above for the true raw source)",
      "url": "derived from https://www150.statcan.gc.ca/n1/tbl/csv/17100005-eng.zip",
      "retrieved_utc": "2026-09-13T01:37:38Z",
      "raw_sha256": "0c68fc3fa66f763c164f37d952c17473f3f3d814a49660776faa1f0d69cd99b4",
      "raw_bytes": 3595,
      "notes": "GEO=Canada, Gender='Total - gender', Age group in {13..19 years}, REF_DATE 2020-2024. 35 data rows, kept in place of the 38.8MB full uncompressed CSV."
     }
    }
   },
   {
    "file": "data/validation/gbr_jpn_aus.provenance.json",
    "record": {
     "GBR": {
      "url": "https://www.ons.gov.uk/file?uri=/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland/mid2011tomid2024/myebtablesuk20112024.xlsx",
      "source_landing_page": "https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernireland",
      "retrieved_utc": "2026-09-13T01:35:00Z",
      "raw_file": "raw/ons_myebtablesuk20112024.xlsx",
      "raw_sha256": "321f27261c9cf110ea44ca10d198580ec3969bc1171c780cba5596184df9b04d",
      "raw_bytes": 47036552,
      "notes": "ONS 'Mid-2011 to mid-2024 detailed time series' back-series workbook (mid2011tomid2024/myebtablesuk20112024.xlsx). Sheet MYEB4 = 'Annual mid-year population estimates for United Kingdom and constituent countries, by sex and single year of age'. Filtered to Code=K02000001 (UNITED KINGDOM), summed sex m+f to get persons, ages 13-19, years 2022-2024 (population_2022..population_2024 columns). Per source Notes sheet: 'population_20xx: Estimated mid-year population e.g. where population_2018 is the estimate for 30 June 2018' -> reference_date = 30 June of the stated year. Estimates are unrounded as published (not accurate to single-person level per ONS caveat). Mid-2022/2023 England & Wales figures were later revised (see source Notes item 6); Scotland mid-2011-2023 also revised (item 7) -- this file reflects those revisions since it is the current back-series release as of retrieval date, not the original point-in-time mid-year bulletins."
     },
     "JPN": {
      "url_2024": "https://www.e-stat.go.jp/en/stat-search/file-download?statInfId=000040268910&fileKind=4",
      "url_2023": "https://www.e-stat.go.jp/en/stat-search/file-download?statInfId=000040166025&fileKind=4",
      "url_2022": "https://www.e-stat.go.jp/en/stat-search/file-download?statInfId=000040045487&fileKind=4",
      "source_landing_pages": [
       "https://www.stat.go.jp/english/data/jinsui/2024np/index.html",
       "https://www.stat.go.jp/english/data/jinsui/2023np/index.html",
       "https://www.stat.go.jp/english/data/jinsui/2022np/index.html"
      ],
      "retrieved_utc": "2026-09-13T01:37:00Z",
      "raw_files": {
       "2024": "raw/jpn_table1_2024.xlsx",
       "2023": "raw/jpn_table1_2023.xlsx",
       "2022": "raw/jpn_table1_2022.xlsx"
      },
      "raw_sha256": {
       "2024": "4d0596faf9a6bf6dc48179c90efe7e1a4eeee04580612be5e0bb58a5076223e2",
       "2023": "dda3c0cd31c837f307e67c44c39f865804ced974feee1bb7c1825aa49172baf8",
       "2022": "10859c35ba6dc35215de3a81d32ae70daba7078afeb16760c8186a4db8608271"
      },
      "raw_bytes": {
       "2024": 23495,
       "2023": 23482,
       "2022": 23436
      },
      "notes": "Japan Statistics Bureau 'Current Population Estimates as of October 1, <year>', Table 1 'Population by Age (Single Years), Sex and Sex ratio', 'Both sexes' / Total population column, downloaded as the underlying xlsx from the e-Stat file-download endpoint linked off each year's stat.go.jp/english/data/jinsui/<year>np/index.html page (the stat.go.jp page itself only shows a summary; the actual table workbook is hosted on e-stat.go.jp, statInfId identified per year via https://www.e-stat.go.jp/en/stat-search/files?lid=<lid>). IMPORTANT CAVEAT: the source table reports population in whole thousands ('単位 千人' / 'Unit: thousand persons'), i.e. each single-year-of-age figure is already rounded to the nearest 1,000 by the Statistics Bureau -- this is a genuine precision limit of the published table, not an error introduced here. population = published_thousands * 1000. reference_date = October 1 of the stated year (Japan's population estimates are anchored to Oct 1, not a mid-year date)."
     },
     "AUS": {
      "url": "https://www.abs.gov.au/statistics/people/population/national-state-and-territory-population/dec-2025/3101059.xlsx",
      "source_landing_page": "https://www.abs.gov.au/statistics/people/population/national-state-and-territory-population/latest-release",
      "retrieved_utc": "2026-09-13T01:39:00Z",
      "raw_file": "raw/abs_3101059.xlsx",
      "raw_sha256": "7bb703d04bc08c7eb488f617c863e3c536ef8f95980aa048b5d52c360d1942c3",
      "raw_bytes": 201862,
      "notes": "ABS cat. 3101.0 'National, state and territory population', TABLE 59: 'Estimated Resident Population By Single Year Of Age, Australia' time-series workbook, linked from the Dec-2025 quarterly release landing page (single-year-of-age series are annual, series end 30 June 2025 as of this release, current back-series updated with each quarterly ERP release incl. any revisions). Used the 'Persons' (both sexes) series IDs for ages 13-19 (A2159135V..A2159141R) from sheets Data1/Data2, years 2023-2025 (row keyed by date 1-Jun-<year>, which ABS uses as the internal row label for the 30-June reference date). reference_date = 30 June of the stated year, per ABS convention for Estimated Resident Population (ERP)."
     },
     "general_notes": "All three sources are the current, most-recently-published back series as of retrieval (2026-09-13), not archived point-in-time releases -- so historical years reflect any subsequent ABS/ONS/Stat.go.jp revisions, consistent with each agency's own published back series. All figures are independent official national statistics office series and were not derived from or cross-checked against the project's WPP2024-based teen population pipeline."
    }
   }
  ]
 },
 "uncertainty": {
  "status": "not_published",
  "reason": "A sensitivity/model-spread interval has not yet passed validation. The independent validation above is the accuracy evidence for this release."
 },
 "downloads": {
  "latest.csv": "Estimated population aged 13-19 per entity at the bake instant",
  "latest.json": "Same, with the global total and metadata",
  "seasonality_profiles.csv": "Monthly birth shares per entity with source tier and years used",
  "model_vs_national.csv": "Model versus independent national statistics, per country-year",
  "births_by_month_sources.csv": "All sourced monthly birth counts used (HFD, Eurostat, UN DYB, published)",
  "model-release.json": "This record",
  "LICENSE.txt": "CC BY 4.0"
 },
 "license": "CC BY 4.0"
}