HistoryData
HistoryData is a reference database for world history. It catalogs people, events, wars, empires, places, and the economic and demographic indicators that frame them — from antiquity to the present — as structured, cross-linked entries drawn from primary sources.
What the database covers
Coverage spans from roughly 3000 BC to the present day — Pharaonic dynasties, classical antiquity, the medieval courts of Eurasia, colonial empires, industrial-era conflicts, and the modern nation-state. The core unit is the historical entity, not the article: every person, event, war, empire, and place carries discrete metadata (dates, coordinates, participants, casualties, outcomes) that can be filtered, ranked, and cross-referenced.
Twelve verticals link to each other through shared participants, locations, and time windows. A war links to its commanders; each commander links back to every conflict they appear in. A period ribbon places an empire in context with its predecessors and successors. A place page surfaces the wars, treaties, and notable residents tied to those coordinates. The goal is a graph dense enough that browsing one entry reliably uncovers three more worth reading.
Alongside the historical entities, the site publishes long-run quantitative series — GDP, inflation, life expectancy, CO₂ emissions, corruption index, average adult height, and the full UN population pyramids (1950–2100) — for 193 countries. These are kept separate from editorial content but cross-linked from country and person pages where relevant.
Data methodology
SPARQL queries against Wikidata pull structured facts — birth and death dates, coordinates, citizenship, occupations, reign spans, casualty figures — directly from community-maintained graphs.
Narrative fields (biographies, war outlines, empire arcs) are generated by Claude Sonnet against strict schemas and a 200 records-per-day budget. Output is validated before it reaches the database.
Each generated entry is reconciled against the Wikipedia article for the same entity. Conflicting dates, misassigned nationalities, and broken references fail validation and are queued for review.
Entries are rendered as static pages through Next.js ISR, indexed in XML sitemaps, and released under CC BY-SA 4.0. The full catalog regenerates monthly.
Primary data sources
Editorial standards
Facts over narrative
The unit of record is a verified fact — a date, a coordinate, a name, a count. Interpretive claims about cause and meaning belong in scholarship, not in a reference index. We catalog; we do not argue.
Ranges when sources disagree
Historical counts — casualties, populations, reigns — rarely converge to a single number. Where scholarly estimates diverge, we publish the range and note the source rather than choosing a headline figure.
Corrections welcomed
Mistakes reach production; 2,106 misassigned nationalities were corrected in a single pass in 2026. Email contact@historydata.com with specific corrections and we will publish the fix.
AI-assisted enrichment runs nightly under a 200-persons-per-day budget, a guardrail added after an uncapped pass in 2026 exhausted credits. At that pace, full coverage lands in approximately 2.7 years. Height data is sourced separately from Wikidata P2048 and covers 9,372 persons.