About Unequal World
Unequal World is a living atlas of global inequality that holds three things as equals: original Unequal Scenes aerial photography, frontier neuroscience on how environments shape brain aging, and authoritative global data. The ambition is one place to see, compare and verify how unequal the world has become, and to feel it.
The photography is Johnny Miller's Unequal Scenes: the aerial archive that makes spatial inequality impossible to look away from, here geolocated to the cities and the data it portrays. The science is Brain age, our estimate of how a place's environment (air, water, green space, inequality, infrastructure) is associated with accelerated brain aging, grounded in Legaz et al. (2026, Nature Medicine) and developed as a research collaboration (not a formal endorsement). It is a directional, population-level index, not a diagnosis; the full disclaimer and institutional partners are set out on the Methodology page.
The data is drawn from the World Bank (Data360 / WDI), IMF and UNICEF, the World Inequality Database (WID.world), ESA and NASA satellites, and national census offices across 20 countries; every country indicator is geolocated, time-aware, and one click from its authoritative source. On top of it we build city-level maps ourselves: 100 m "development burden" grids and 1930s redlining boundaries laid over today's census income. The whole platform was built end-to-end with AI.
The four lenses
- Health: leads with Brain age, our derived estimate of environmental brain aging (Legaz et al. 2026, Nature Medicine), plus life expectancy, water, literacy.
- Economy: Gini, poverty, GDP/capita, internet; toggle financial flows for animated ODA / remittances / FDI / climate-debt arcs.
- Urbanization: urban population, slum share, electricity, internet access.
- Planet: CO₂ per capita, forest area, renewable electricity.
How the analysis works (and where AI fits)
- Algorithmic cross-referencing: deterministic statistics (not machine learning) that read across the World Bank series to surface source-sealed comparisons a single-indicator chart can't: peer anomalies (click any country → "how it defies its income + region peers", robust median + MAD, sample size disclosed) and divergences (teal "↑↓" pins on the time-slider globe, where two indicators that usually move together broke apart; click one to graph both series). Correlation, not causation; every claim carries a Data360 verify chip.
- Inflection detection: an algorithm scans every World Bank time-series for statistically significant trend reversals (5-year rolling slope change > 1.5σ). Detected bends are ranked by severity (bend size in σ, full reversals and wrong-direction bends weighted up), and only those that carry a hand-researched, source-linked cause (or map to a known global event) are shown, colour-coded improvement / decline / catastrophe; the unexplained tail is hidden.
- Generative AI: used to build the tool: one person assembled it with generative AI as the primary engine, from satellite data, national census offices across 20 countries, and the World Bank API. The runtime analytics above are deterministic, not AI.
Aligned to Data360's five focus areas
Data360 organises its thousands of indicators into five areas. Our tabs map directly onto them, so the platform reads as a focused lens on the World Bank's own taxonomy:
- People → our Health tab (life expectancy, water, literacy, brain age).
- Prosperity → our Economy tab (Gini, poverty, GDP/capita, flows).
- Infrastructure → our Urbanization tab (urban & slum share, electricity).
- Planet → our Planet tab (CO₂, forest, renewable electricity).
- Digital → partially covered today (internet access on Economy & Urbanization); a dedicated Digital tab (broadband, mobile) is the next tab on the roadmap.
Data & provenance
- World Bank Data360 / WDI: every country indicator; one-click "Verify on Data360" on each value.
- Federated Data360 sources: beyond WDI, the country panel pulls real values from IMF (World Economic Outlook, current account) and UNICEF (child underweight), proving cross-database reach. Each carries its own Data360 verify chip.
- World Inequality Database (WID.world): top 1% and top 10% income and wealth shares (Piketty, Saez, Zucman, Chancel), shown as Economy lenses beyond headline Gini; each verifies to wid.world.
- Satellite: ESA Sentinel-2 (vegetation), Sentinel-5P (NO₂), NASA VIIRS (night lights), EU JRC GHSL (built-up + population).
- Neighbourhood wealth: national census & statistics offices across 20 countries (IBGE census sector, US ACS block group, StatsSA Small Area Layer, INEI Peru, CONAPO Mexico, DANE Colombia, etc.).
- Future: IIASA Shared Socioeconomic Pathways (post-2024 projections).
How to trust each number
Every value is tiered so you know what you are reading: measured data is shown as-is and source-linked, anything we derive (the development-burden composite, Brain age) is labelled derived, and known measurement artifacts are flagged or suppressed rather than shown as events.
The tier system, the comparability rules, the burden formula, satellite biases and every known limitation are documented on the Methodology page.