← All datasets

Every War by Death Toll

A CodeSights flagship for the history and conflict-data crowd: 132 of history's deadliest wars with their death-toll range (low and high, unit-parsed from "70-85 million" style text), start and end years, combatants, location and, for the deadliest 33, the share of the living they killed. Parsed from the single ranked table of Wikipedia's list, read from rendered HTML because the wikitext cells shatter on multi-line citation templates. The famous finding is a rank reversal, the same lesson as Every Pandemic in a new domain: by absolute deaths World War II leads at 70 to 85 million, roughly three times the Taiping Rebellion behind it, but by share of humanity the An Lushan rebellion took 6.04 percent on 13 million dead, nearly double WWII's 3.04 to 3.69 percent, and the Three Kingdoms wars took up to 17.89 percent, by test. One sort by each column reorders the carnage. The deadliest wars span from the Greco-Persian Wars of the fifth century BCE (negative start years survive the parse) to eleven conflicts still ongoing, and the 20th century holds 51 of the 132, more than any other. Death bounds are ordered low to high by test, every war carries a Wikidata QID for joining, and the share column is kept as stated rather than invented. A different grain from the Battles topic: a war, not an event inside one. Eleven acceptance tests anchored on World War II as the absolute maximum and An Lushan as the share champion that outranks it.
132 rows 1 file last pulled 2026-10-15
Source: https://en.wikipedia.org/wiki/List_of_wars_by_death_toll
Wars 132 rows × 13 columns
WarWar_IDStart_YearEnd_YearDuration_YearsCombatantsLocationDeaths_LowDeaths_HighDeaths_RawPct_Global_PopulationOngoingWar_URL

Questions and answers

What is the Every War by Death Toll dataset?

A CodeSights flagship for the history and conflict-data crowd: 132 of history's deadliest wars with their death-toll range (low and high, unit-parsed from "70-85 million" style text), start and end years, combatants, location and, for the deadliest 33, the share of the living they killed. Parsed from the single ranked table of Wikipedia's list, read from rendered HTML because the wikitext cells shatter on multi-line citation templates. The famous finding is a rank reversal, the same lesson as Every Pandemic in a new domain: by absolute deaths World War II leads at 70 to 85 million, roughly three times the Taiping Rebellion behind it, but by share of humanity the An Lushan rebellion took 6.04 percent on 13 million dead, nearly double WWII's 3.04 to 3.69 percent, and the Three Kingdoms wars took up to 17.89 percent, by test. One sort by each column reorders the carnage. The deadliest wars span from the Greco-Persian Wars of the fifth century BCE (negative start years survive the parse) to eleven conflicts still ongoing, and the 20th century holds 51 of the 132, more than any other. Death bounds are ordered low to high by test, every war carries a Wikidata QID for joining, and the share column is kept as stated rather than invented. A different grain from the Battles topic: a war, not an event inside one. Eleven acceptance tests anchored on World War II as the absolute maximum and An Lushan as the share champion that outranks it.

How big is the Every War by Death Toll dataset?

132 rows in one CSV file, with 13 columns in total: War, War_ID, Start_Year, End_Year, Duration_Years, Combatants, Location, Deaths_Low, Deaths_High, Deaths_Raw, Pct_Global_Population, Ongoing, War_URL.

Where does the data come from?

It is built by programmatically scraping en.wikipedia.org, last pulled on 2026-10-15. Datasets are versioned; older versions stay downloadable.

Is the dataset free to download?

Yes. Every CodeSights dataset is completely free as a CSV download; a free account is all it takes. Anyone can preview the data without signing in.

Can I see the code that built this dataset?

Yes. The exact Python scraper that built it is viewable on the dataset page by any signed-in member, so every number is reproducible.

Can the data contain errors?

Automated scraping leaves room for error and the underlying sources change over time, so no version is guaranteed accurate or complete. If a number matters, verify it against the original source.

Browse every free dataset on CodeSights