Nine processes regulate the stability, resilience, and life-support functions of the Earth system. Seven have now been breached, ocean acidification for the first time in 2025. This is the framework within which the datasets below sit.
Nine Earth system processes regulate planetary stability, resilience, and life-support functions, each tracked by one or two control variables. Wedge length shows how far each has moved beyond the safe operating space; softened edges mark variables whose current value or risk zone is not precisely quantified. Values from the Planetary Health Check 2025, Potsdam Institute for Climate Impact Research. DOI 10.48485/pik.2025.017, CC BY 4.0.
Full page: the nine boundaries explained, and what crossing them means →
Every country either draws more from nature than its ecosystems can regenerate, or holds a reserve. This map tracks that balance across 130 countries, alongside footprint and biocapacity in both per-person and total terms.
National Footprint and Biocapacity Accounts, 2025. Countries in grey have insufficient or inconclusive data. Source: Global Footprint Network National Footprint and Biocapacity Accounts, via the Open Data Platform.
Full page: methodology, how to read it, and related research →
The SDG Index scores countries from 0 to 100 on their progress across all 17 Sustainable Development Goals, where 100 means every Goal has been achieved. Compiled annually by the Sustainable Development Solutions Network.
SDG Index 2026. Scores run 0–100, where 100 means every Goal has been achieved. Countries in grey are not included in the Index. Source: Sachs, J.D., Lafortune, G., Fuller, G., Iablonovski, G. (2026). Implementing Sustainable Development: 2030 and Beyond. Sustainable Development Report 2026. Paris: SDSN, Dublin: Dublin University Press. DOI: 10.82163/225
Full page: methodology, how to read the Index, and related research →
Health, education, and income combined into a single value between 0 and 1, across 184 countries from 1990 to 2023. Read alongside the ecological footprint map, it shows what each country’s development has cost in resource terms.
The Human Development Index combines life expectancy at birth, expected and mean years of schooling, and gross national income per capita into a single value between 0 and 1. Source: UNDP Human Development Report, via Our World in Data. UNDP classifies scores above 0.800 as very high human development, 0.700 to 0.799 as high, 0.550 to 0.699 as medium, and below 0.550 as low.
Full page: what the Index combines, and what it leaves out →
Adult, female and youth literacy, secondary and tertiary enrolment, and public spending on education across 187 countries. Literacy is measured by census rather than annually, so those views show each country’s most recent reading with the year on hover — Morocco’s is from 2014. Rich countries stop reporting literacy once it nears universal, which is why they appear blank rather than at 100%.
Source: UNESCO Institute for Statistics and World Bank. Literacy is measured by census or household survey rather than annually, so each country shows its most recent reading with the year given on hover — some are a decade old. High-income countries stop reporting literacy once it approaches universal, which is why much of Europe, North America and Japan is blank on those views rather than at 100%. Enrolment ratios are gross, counting all enrolled pupils against the population of the relevant age group, so they can exceed 100% where students repeat years or enrol late.
Extreme poverty at the international $3 a day line, alongside each country’s own national poverty line. The gap between the two is often larger than the gap between countries — India records 5.3% by the international measure and 21.9% by its own.
Source: World Bank Poverty and Inequality Platform, via Our World in Data. Each country shows its most recent available observation, with the year given on hover. Poverty data comes from household surveys conducted irregularly, so reference years differ between countries and direct comparison carries that caveat. The international line is set at $3 a day in 2021 purchasing power parity.
Income inequality, safely managed water and sanitation, and out-of-pocket health costs across 187 countries. The Gini index is the standard inequality measure and communicates almost nothing; the income shares beside it say the same thing in terms anyone can picture, and the hover panel divides one by the other to give the ratio per person.
Source: World Bank World Development Indicators and WHO/UNICEF Joint Monitoring Programme. Inequality is measured from household surveys run every few years, so those three views show each country's most recent reading with the year on hover — some are a decade old. The Gini index runs from 0, where everyone has the same income, to 100, where one person has it all; real values fall between roughly 24 and 63. Income shares are often clearer: what the richest tenth and the poorest fifth actually receive. "Safely managed" water and sanitation are stricter standards than "basic" — they require an improved source on the premises, available when needed and free from contamination. Out-of-pocket health spending is the share of health costs paid directly by households at the point of care, and it is the measure most closely tied to illness pushing people into poverty.
Undernourishment, severe food insecurity, the cost of a healthy diet and calorie supply across 186 countries. Severe food insecurity is the measure that asks people directly whether they have gone without food — South Sudan records 63.1%, Italy 0.5%. Adequate nutrition costs between $2.50 and $7.29 a day depending on local prices; what varies far more is the income available to pay it.
Source: FAO Suite of Food Security Indicators, via Our World in Data. Undernourishment estimates the share of the population whose habitual food consumption is insufficient for a normal active life; it is modelled from food supply and distribution data rather than measured directly. Severe food insecurity comes from the Food Insecurity Experience Scale, which asks people whether they have run out of food or gone a whole day without eating. The cost of a healthy diet is the least expensive combination of local foods meeting dietary guidelines, in purchasing-power-adjusted dollars per person per day — set it against the $3 a day extreme poverty line. Calorie supply measures what is available in the food system, not what people eat, and takes no account of distribution.
Diabetes prevalence, adult obesity and sugar supply across 188 countries. The temptation is to read a chain from sugar to obesity to diabetes, and the data does not support it — several high-sugar countries have modest diabetes rates while parts of South Asia record severe prevalence on unremarkable sugar supply. Obesity is the one continuous series, running from 1980, and its trend is the most dramatic on this page.
Source: IDF Diabetes Atlas, WHO Global Health Observatory and FAO food balance sheets, via Our World in Data. Obesity runs as an annual series from 1980. Diabetes and sugar are published at intervals rather than yearly, so those views show each country's most recent reading with the year on hover. Sugar is supply, not consumption — calories entering a country's food system per person per day, before any account of waste, which means it overstates what people actually eat, probably by a fifth or more. Diabetes prevalence covers ages 20 to 79 and includes both type 1 and type 2, though type 2 accounts for the large majority and is the type associated with diet.
GDP per capita, total GDP, trade openness, inflation and unemployment across 188 countries from 1990. Income is the strongest single predictor of almost every other indicator on this page — keeping it here makes that relationship checkable rather than assumed.
Source: World Bank World Development Indicators, via Our World in Data. GDP figures are in international dollars at purchasing power parity, which adjusts for price differences between countries and is the appropriate basis for comparison. Trade openness is exports plus imports as a share of GDP, so it exceeds 100% where re-exports are large. Values shown at five-year intervals from 1990, with annual data from 2019.
Full page: why economic data sits alongside the environmental →
Seven macro indicators across 186 countries from 2000 — GDP growth, inflation, real interest rates, current account, fiscal balance, reserves, and currency movement against the dollar. Hovering any country shows all seven at once, which is the point: a single indicator rarely tells you much, and the combination is what analysts actually read.
Source: World Bank World Development Indicators and IMF International Financial Statistics. Coverage varies by measure: GDP growth and inflation reach most countries, while fiscal balance and real interest rates are reported by fewer than half, since not every government publishes them on a comparable basis. Currency change is calculated against the US dollar, where positive means the currency strengthened. Reserves are expressed in months of import cover, which is the standard adequacy measure — three months is the conventional floor.
Resource rents as a share of GDP across 186 countries, broken down by oil, gas, coal, minerals and forest. A rent is the surplus between what a commodity sells for and what it costs to extract — so this measures how much of an economy consists of drawing down natural capital rather than producing anything. Read alongside the governance map, the resource curse is visible and not subtle.
Source: World Bank World Development Indicators. A resource rent is the difference between what a commodity sells for and what it costs to extract — the surplus available to whoever controls the resource. Expressed as a share of GDP, it measures how much of an economy consists of drawing down natural capital rather than producing goods and services. The five components sum to the total. Figures track commodity prices as much as extraction volumes, so a country's rents can halve in a year without a single well closing. The series ends in 2021.
Six dimensions of institutional quality — government effectiveness, regulatory quality, rule of law, control of corruption, voice and accountability, and political stability — across 188 countries from 1996. Environmental policy does not implement itself, and these are the closest available proxy for the institutional capacity it requires.
Source: Worldwide Governance Indicators, World Bank. Each indicator runs from roughly −2.5 to +2.5, with zero at the world average for the year. The scores are aggregates of perception data from around 35 survey and expert sources, so they measure assessed institutional quality rather than directly observed outcomes. Standard errors are substantial, and small differences between countries should not be treated as meaningful.
Full page: what the six dimensions measure, and how not to read them →
Recorded deaths in armed conflicts across 185 countries, 1989 to 2025, by where the fighting occurred. These are documented deaths — UCDP counts only what it can verify from reported sources, so the real figures are higher. The hover panel shows UCDP’s own low and high estimates alongside the central figure, and the distance between them says a good deal about how contested the reporting is.
Source: Uppsala Conflict Data Program, via Our World in Data. Deaths are recorded by where the fighting occurred, not by the nationality of those killed. UCDP counts only deaths it can verify from reported sources, so these are documented deaths rather than total deaths — the true figures are higher, often substantially. The low and high estimates are UCDP's own uncertainty range, and the distance between them is a useful indicator of how contested the reporting is. Countries shown in the palest band recorded no conflict deaths that year, which is different from having no data.
Homicide rates, female victim rates and prison population across 183 countries from 1990. There is no general crime rate here deliberately: recorded crime is not comparable between countries, since it tracks reporting culture and police practice as much as incidence. Homicide is the exception, and the female victim rate behaves differently from the aggregate because women are far more often killed at home.
Source: UNODC and the World Prison Brief, via Our World in Data. Homicide is the only crime statistic comparable across countries — a death is recorded whatever the reporting culture, whereas theft, assault and sexual violence figures track willingness to report and police recording practice at least as much as underlying incidence. That is why no general crime rate appears here: recorded crime is not comparable between countries, and a map of it would produce confident colour where no comparison exists. UNODC reports with a two to three year lag, so the most recent years are thinly covered. Prison population has no common reporting year at all — each country is updated whenever its national statistics appear — so that view shows each country's most recent reading, with the year given on hover. It measures state action rather than criminality, and the two are less related than intuition suggests.
Internet use, mobile subscriptions and fixed broadband across 188 countries from 1995. Switching between mobile and broadband reorders the map substantially — much of the world came online without ever building fixed-line infrastructure, and that difference shapes what connectivity is actually good for.
Source: International Telecommunication Union and World Bank, via Our World in Data. Internet use counts anyone who has used the internet in the previous three months, by any device. Mobile subscriptions can exceed 100 per hundred people because many people hold more than one SIM. Fixed broadband counts household connections rather than individuals, so it understates access where one connection serves several people.
Full page: why mobile and broadband tell different stories →
Eleven demographic indicators across 182 countries from 1960 to 2023 — population, growth rate, median age, fertility, age structure, life expectancy, urbanisation and density. Demography sits underneath most of the other datasets here, since population is the multiplier on every per-capita figure.
Source: UN World Population Prospects and World Bank, via Our World in Data. Values shown at five-year intervals from 1960, with annual data for 2020 onward, ending in 2023 where the underlying series currently stops. Sex ratio at birth is biologically about 105 males per 100 females; values meaningfully above that indicate sex-selective practices.
Female labour force participation and the gender participation gap across 174 countries from 1990. The map opens on the gap rather than the female rate, because participation is often highest in the poorest countries — where subsistence work counts as employment and not working is not an option. Morocco records a gap of 48.5 percentage points; Norway 7.1.
Source: International Labour Organization modelled estimates, via the World Bank. Participation counts anyone aged 15 or over who is working or actively seeking work. The map opens on the participation gap — male minus female participation in percentage points — because the female rate alone is easy to misread: it is often high in the poorest countries, where subsistence agriculture counts as employment and not working is not an option. The gap is the clearer measure of exclusion.
The number of valid certificates to ISO management system standards, reported by country. Switch between quality, environmental, occupational health and safety, food safety, and energy management standards.
Valid certificates as at 31 December 2024. Source: The ISO Survey of Management System Standard Certifications, ISO and the International Accreditation Forum. Participation by certification bodies is voluntary and varies between years, so country-level changes may reflect reporting coverage rather than genuine shifts in certification. The survey is not a register of certificates.
Coal consumption across 79 countries from 1990 to 2025. Switch between total consumption, consumption per person, and coal’s share of primary energy — each answers a different question about dependence and responsibility.
Coal consumption covers commercial solid fuels only: hard coal, lignite, and other solid fuels. It excludes coal converted to liquid or gaseous fuels but includes coal consumed in transformation processes such as electricity generation. Source: Energy Institute Statistical Review, via Our World in Data. Coverage is limited to the countries surveyed by the Energy Institute, so many smaller economies are absent rather than at zero.
Production and consumption-based CO2, plus all greenhouse gases, across 187 countries from 1960. The consumption-based view is the one worth switching to: it reassigns the emissions of imported goods to the country that buys them, and several wealthy economies look considerably less improved once it does.
Source: Global Carbon Budget and Climate Watch, via Our World in Data. Production-based emissions count what is emitted inside a country's borders; consumption-based emissions reassign those of imported goods to the country that buys them, and subtract the emissions of exports. The difference matters: several wealthy economies have reduced territorial emissions partly by importing what they once manufactured. Consumption-based figures require trade-flow modelling and exist for 120 countries only, to 2023. Greenhouse gas totals include methane, nitrous oxide and land use change alongside CO2, expressed in CO2 equivalents.
Each source as a share of primary energy consumption — coal, oil, gas, nuclear, hydro, wind and solar — across 79 countries from 1990 to 2025. Primary energy covers transport and heating as well as electricity, which is why renewable shares look smaller here than in the headlines.
Each source as a percentage of primary energy consumption, covering all energy use including transport, heating and industry rather than electricity alone. Source: Energy Institute Statistical Review, via Our World in Data. Nuclear is derived as low-carbon energy minus renewables. Traditional biomass is excluded, which understates energy use in lower-income countries. Coverage is limited to the countries the Energy Institute surveys.
Refining capacity, throughput and utilisation across 74 countries from 1980. Crude oil is not a usable fuel — refining capacity determines whether a country can meet its own fuel demand or must import refined products. Morocco has reported none since its Mohammedia refinery closed in 2015.
Source: Energy Institute Statistical Review of World Energy. Capacity is atmospheric distillation capacity at year end on a calendar-day basis. Utilisation can exceed 100% because refineries routinely operate above nameplate capacity. Coverage is limited to the countries the Energy Institute surveys; a value of zero means a country reported no operating refining capacity, which is different from being absent from the data.
Electric car sales and stock across 56 markets from 2010. Norway sells 97% electric but only 36% of its cars on the road are — the gap between those two figures is how long a vehicle fleet takes to turn over, and it is the number that matters for emissions.
Source: IEA Global EV Outlook, via Our World in Data. Electric cars include both battery electric and plug-in hybrid vehicles. Coverage is limited to the roughly fifty markets the IEA tracks, so most of Africa, the Middle East and Central Asia are absent from the data rather than at zero. The gap between sales share and fleet share is the important one: it measures how long a vehicle fleet takes to turn over even where new sales have shifted almost entirely.
Freshwater withdrawal as a share of available renewable resources, from 2000 to 2023. Values above 100% mean a country withdraws more than its renewable supply, drawing down aquifers or relying on desalination.
SDG indicator 6.4.2, the ratio of freshwater withdrawn to available renewable freshwater resources after environmental flow requirements. Values above 100% indicate withdrawal exceeding renewable supply, sustained by aquifer depletion or desalination. Source: FAO AQUASTAT, via the UN SDG Global Database. National figures average across a country, so they can understate severe scarcity within particular basins or regions.
Full page: how water stress is measured, and related research →
Withdrawal split between agriculture, industry and households, alongside renewable resources per person, across 183 countries. Agriculture takes roughly 70% of the world’s freshwater and above 85% across South Asia and North Africa — which means water policy is mostly agricultural policy, however much the public campaigns focus on household use.
Source: FAO AQUASTAT, via the World Bank. The three sectoral shares sum to 100% of each country's total withdrawal. Agriculture dominates almost everywhere, which is why water policy is largely agricultural policy. Internal renewable resources count only water generated within a country's own territory from rainfall — river flow arriving from upstream neighbours is excluded, so countries dependent on transboundary rivers appear far drier here than their actual supply suggests. AQUASTAT figures are updated irregularly and many countries report once every few years.
Mean annual exposure to PM2.5 across 171 countries from 1990. The WHO guideline is 5 micrograms per cubic metre; Morocco sits at 18.7, China at 32, India at 54. Almost no country meets the standard, which is why the WHO publishes four interim targets rather than one.
Mean annual population-weighted exposure to fine particulate matter under 2.5 micrometres, in micrograms per cubic metre. Source: World Bank World Development Indicators. The WHO air quality guideline is 5 µg/m³; the interim targets run at 35, 25, 15 and 10. National averages weight by where people live, but still conceal wide variation between cities and rural areas within a country.
Full page: what PM2.5 is, and why national averages hide the cities →
Forest cover, net change, deforestation and satellite-observed tree cover loss across 168 countries. These four measure different things and are routinely conflated — a country can show net forest gain while deforestation continues, if planting outpaces clearance.
Source: FAO Global Forest Resources Assessment and Global Forest Watch, via Our World in Data. Tree cover loss is satellite-observed and reported annually from 2001; it counts all canopy loss including commercial harvest and fire, much of which regrows. Deforestation is the permanent conversion of forest to another land use, reported by countries to the FAO only in five-year assessment periods, which is why its year buttons are limited. The two measure different things and tree cover loss is always the larger.
Kilograms of food wasted per person per year across household, food service, and retail, in 184 countries. Read the full page before drawing country comparisons — most of these figures are modelled rather than measured.
Kilograms of food wasted per person per year, 2022 estimates. Source: UNEP Food Waste Index Report, via Our World in Data. Most countries have no direct measurement: household figures are estimated by extrapolating from countries with primary studies, adjusted for income level, while retail and food service figures are largely modelled defaults applied by income group. Identical values across unrelated countries indicate modelled rather than measured data.
Recorded natural hazard events from 2000 to 2025 across 180 countries, from the EM-DAT international disaster database. Switch between hazard types and between events, deaths, and people affected.
Natural hazard events recorded 2000-2025, aggregated by country. Source: EM-DAT, CRED / Universite catholique de Louvain. EM-DAT records events meeting at least one threshold: ten or more deaths, a hundred or more people affected, a declaration of emergency, or a call for international assistance. Reporting quality varies between countries and hazard types, so differences partly reflect monitoring capacity rather than exposure alone. People affected is generally a better guide to impact than event counts, which favour large and well-monitored countries.
Full page: what EM-DAT records, and why reporting bias matters →
How people rate their own lives on a scale from 0 to 10, across 162 countries from 2011 to 2025. Drawn from the Gallup World Poll and published annually in the World Happiness Report.
Average response to the Cantril ladder question, which asks people to rate their present life on a scale from 0, the worst possible life, to 10, the best possible. Roughly 1,000 respondents per country each year, from the Gallup World Poll. Source: World Happiness Report, via Our World in Data. Self-reported life evaluation captures something different from income or health alone, which is why it is used as a complement to GDP rather than a substitute for it.
Full page: the Cantril ladder, and what life evaluation captures →
Documents, citations, h-index and self-citation rates across 204 countries, cumulative 1996 to 2025. Switching between volume and citations per document reorders the map almost entirely — the first measures scale, the second measures reception, and they are routinely conflated.
Source: SCImago Journal & Country Rank, based on Scopus data. Figures are cumulative over 1996 to 2025, not annual. Documents are attributed to a country when at least one author holds an affiliation there, so internationally co-authored papers count for every country involved and national totals sum to more than the world total. Scopus coverage is uneven by language and discipline, favouring English-language journals and the natural sciences.
Scholarly publications on artificial intelligence across 169 countries, 2016 to 2024. China leads on volume with 116,000 a year; Singapore leads on intensity at 765 per million, nine times China’s rate. The map opens on per-capita, because the total view is largely a map of population.
Source: Stanford AI Index Report, via Our World in Data. Counts are annual, not cumulative. Publications are attributed by author affiliation, so internationally co-authored work counts for every country involved. What qualifies as an AI publication is defined by keyword and subject classification, which has widened as the field has, so growth partly reflects definitional drift as well as real expansion.
Full page: scale, intensity, and what publication counts miss →