How each number reaches the page

SDG indicator 8.7.1 (SL_TLF_CHLDEA), both sexes. Retrieved by direct download of the Our World in Data grapher CSV (children-aged-5-17-engaged-in-labor.csv?csvType=full&useColumnShortNames=true, HTTP 200), column _8_7_1__sl_tlf_chldea__5_to_17_years_old__both_sexes; OWID metadata lastUpdated 2026-07-31, nextUpdate 2027-07-31. One value per country: its latest survey year, which varies by country (2010-2025) and is kept beside the value. Engagement threshold varies by age: ages 5-11 at least 1 hour, ages 12-14 at least 14 hours, ages 15-17 at least 43 hours of economic activity in the reference week. Source: International Labour Organization and UNICEF, with minor processing by Our World in Data. Only three-letter ISO country codes are kept: the UN regional aggregate rows (UNSDG_*, UN_*) and OWID_WRL are dropped. 100 countries. Kosovo has no child-labour value in this edition (OWID_KOS carries other indicators only), so it has no cell here; the UNICEF layer still carries it. Zambia and Vietnam move to a newer survey (2023), and their values change with it; the UNICEF Data Warehouse footnote on this indicator says changes in definitions, age groups and questions across surveys affect comparability, so a change between survey years is not a trend.

Pulled live from the UNICEF Data Warehouse SDMX API (sdmx.data.unicef.org, dataflow UNICEF,PT,1.0, indicator PT_CHLD_5-17_LBR_ECON, SEX=_T, AGE=Y5T17, all other breakdowns Total), latest year per country. Survey instrument + year preserved per country in the 'survey' field. This ECON definition is economic activity only (SDG 8.7.1); it excludes the household-chores extension (PT_CHLD_5-17_LBR_ECON-HC).

dol.gov blocked direct fetch (Akamai 403); full-report PDF retrieved via the Georgetown Global Children mirror (globalchildren.georgetown.edu, S3 copy of Full-Report-2024-Findings-on-the-Worst-Forms-of-Child-Labor.pdf) and parsed from the 'TDA Country Assessments, by Assessment' table (pp. 44-49), cross-checked against Figure 2 'Global Breakdown of Country Assessments' and the ICI briefing of Oct 2025. Sub-variant assessments (e.g. 'Minimal Advancement - Efforts Made but Continued Practice that Delayed Advancement', 'No Advancement - Efforts Made but Complicit in Forced Child Labor') are collapsed to their base level. Western Sahara (ESH) is assessed jointly with Morocco in the single row 'Morocco, including Western Sahara' and carries that row's assessment; it is the 131st territory in the report's own count (9+63+46+12+1).

Cumulative 2024 list (11th edition, 204 goods from 82 countries/areas). Extracted from Table 3 "The 2024 TVPRA List by Country" (pp. 19-25) of the official 2024 report PDF (2024-tvpra-list-of-goods.pdf), retrieved via mirror https://www.endchildlabourbd.org/wp-content/uploads/2026/04/2024-tvpra-list-of-goods-1.pdf (dol.gov blocks automated fetches); edition facts cross-checked against Federal Register 89 FR 72897 (doc 2024-20153, govinfo.gov). Type is derived from the Child Labor and Forced Labor columns: CL+FL = good appears in both. Goods in the Forced Child Labor column always also appear in both CL and FL columns (verified) and are therefore covered by CL+FL. Downstream goods made with listed inputs (Table 3 "Downstream" column) are excluded, which omits two downstream-only countries: Netherlands (NLD) and South Korea (KOR).

Direct fetch from state.gov (Tier Placements List section of the 2026 report page; browser-UA GET, HTTP 200; page modified 2026-10-08). No PDF yet: the page says its accessible PDF will be posted once its accessibility updates are complete. The 2026 report covers government efforts from 1 April 2025 to 31 March 2026 and places the same 188 countries and areas as the 2025 edition (Tier 1 35, Tier 2 111, Tier 2 Watch List 18, Tier 3 20, Special Case 4); 21 changed tier. It was released 100 days after its 30 June 2026 statutory deadline. Dominica, Grenada, Sao Tome and Principe and Saint Kitts and Nevis are not fully assessed (the report's own note) and carry no tier. Non-state and sub-state entries ranked by State (Taiwan, Kosovo, Hong Kong, Macau, Aruba, Curaçao, Sint Maarten) carry tipTier with no ILO membership; their c138/c182 fields are false/null by construction, not a ratification finding.

Direct fetch from NORMLEX (both registers: C138 instrument 312283, C182 instrument 312327; browser-UA GET with cookie jar). Declared minimum ages taken from the C138 register's per-country declarations. C138: 177 ratifications, C182: 187 ratifications (universal), all listed In Force.

Retrieved via Our World in Data grapher CSV (share-primary-school-age-out-of-school.csv?csvType=full&useColumnShortNames=true; that chart address now redirects to children-not-in-school?level=primary&metric_type=rate&sex=both, whose CSV gave the same 217 countries and values on 2026-10-15), which re-publishes UNESCO Institute for Statistics (UIS) - Education data with minor processing; column out_of_school_rate_for_children_of_primary_school_age__both_sexes__pct__rofst_1_cp. Counts children of primary school age (typically 6-11 years) not enrolled in ANY level of education, expressed as a percentage of all children in that age group (SDG 4.1.4 out-of-school rate, ROFST_1_CP). One value per country: the latest available year (coverage year varies by country). OWID metadata lastUpdated 2026-05-12, dateDownloaded 2026-07-03, citation 'UNESCO Institute for Statistics (2026)'. Kosovo (XKX) is NOT present in this UIS dataset. OWID regional/income aggregates and World excluded (only ISO3 country codes kept). Education is a recognised DRIVER of child labour and is kept as its own context column, NEVER blended into any score. CAVEATS: reference year varies widely across countries (see year spread); values are model/survey-derived UIS estimates and some countries are missing entirely (no UIS observation). STALE-YEAR FLAG: 18 countries have a latest observation before 2010, including several child-labour-relevant ones (Afghanistan 1993, DR Congo 1999, Haiti 1997, China 1997, Iraq 2007, North Korea 2009, Somalia 1980, Libya 1983) — value is real/published but the reference year is old.

Parsed directly from the GSI 2023 Appendices PDF (cdn.walkfree.org/.../GSI-2023-CH07-Appendices.pdf), Table 26 'Estimated prevalence and number of people in modern slavery, by country'. Underlying figures are the Walk Free / ILO / IOM 2021 Global Estimates of Modern Slavery, national estimates from a risk/imputation model (reference population 2020). 135 of the ~160 GSI countries extracted cleanly; the T-Z tail of the table (incl. Türkiye, Uganda, Ukraine, UAE, UK, USA, Uzbekistan, Venezuela, Vietnam, Yemen, Zambia, Zimbabwe and a few S/T entries) did not extract from the PDF and is parked in the crawl ledger for a manual read. China, India and Pakistan carry national estimates despite survey-data caveats (see the report's data-limitations note).

Downloaded directly from ilo.org with a browser-UA curl (HTTP 200, 2.5 MB, 62-page PDF); text extracted with pypdf. Figures are the 'SDG regions' rows (Total, 5-17 years, %/No.) from the Statistical Tables annex: 'CHILD LABOUR - Percentage and number (in thousands) of children aged 5 to 17 years in child labour' (printed pp. 54-55) and 'HAZARDOUS WORK - Percentage and number (in thousands) of children aged 5 to 17 years in hazardous work' (printed pp. 56-57). Table values are printed in thousands; 'children'/'hazardous' millions here are those thousands divided by 1,000, with the verbatim thousands preserved in childrenThousands/hazardousThousands. The seven SDG-region rows sum exactly to the printed world totals (137,562 and 53,975 thousand).