A CodeSights flagship for the medical and public-health crowd: 261 recorded epidemics and pandemics parsed from the nine chronological and ranked tables of Wikipedia's list, each with its disease, pathogen type (viral, bacterial, parasitic, fungal, prion or an honest unknown for the plagues still debated), start and end years, location, death toll and, for the deadliest, the share of population lost. The death toll is analytics-ready because every figure carries a sortable integer even when the page displays "25-100 million" or "400+". The famous finding is a rank reversal the two columns reveal: by absolute deaths the Black Death (75 million) and the 1918 influenza (58 million) top the list, but by share of population the ancient and colonial plagues dwarf everything modern. The Cocoliztli took up to 80 percent of Mexico and the Black Death 30 to 60 percent of Europe, while COVID-19, for all its 15 million dead, took 0.44 percent of the world, by test. A second story lives in the dates: the count of recorded events per century explodes as surveillance and globalization grow, 88 events since 2000 against 18 before 1500, while deaths per event fall. Pathogen type is derived from the disease name against a closed six-way vocabulary, and bubonic plague, smallpox, yellow fever and cholera dominate the historical record by event count. Years parse BC (the 1350 BC plague of Megiddo), ranges and the present marker, so nine events read as ongoing, HIV/AIDS and COVID-19 among them. Companion to Every Antibiotic: this is what the drugs were fighting. Thirteen acceptance tests anchored on the Black Death, the 1918 flu, COVID-19 and the Plague of Justinian.
261 rows in one CSV file, with 13 columns in total: Epidemic, Epidemic_ID, Disease, Pathogen_Type, Start_Year, End_Year, Duration_Years, Location, Death_Toll, Death_Toll_Raw, Pct_Population_Lost, Ongoing, Epidemic_URL.
It is built by programmatically scraping en.wikipedia.org, last pulled on 2026-10-14. Datasets are versioned; older versions stay downloadable.
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.
Yes. The exact Python scraper that built it is viewable on the dataset page by any signed-in member, so every number is reproducible.
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.