Two joinable datasets covering every human Wikidata records at an Olympic Games: 135,502 athletes and 202,625 participation rows, 1896 to the present, both seasons. The pull fights the graph's real modeling: athletes link participant-in to EVENT items two part-of hops below their Games, OR to the Games directly, and medals live on neither side completely, so the podiums are a hybrid: the event items' winner and ranked-participant statements where they exist, completed by mining each event ARTICLE's infobox medalist rows, with medalists matched by their article URLs and added to the roster when the graph never linked them at all (relay heat swimmers medal too). The proof is the anchor: Michael Phelps ships at exactly 23 gold, 3 silver, 2 bronze, which Wikidata alone had as 15-2-1. Each athlete carries gender, birth and death dates, birthplace, citizenships, the country they competed for, sports, height and weight normalized to cm and kg, Games and event counts, first and last Games with career span, seasons, the full medal ledger with first and last medal years, and the English Wikipedia article (98%). The participation file is the panel: one row per athlete-event with rank and medal where won, plus Games-level rows for athletes the graph places at a Games without saying in what. Mis-linked participations that would make a 101-year-old sprinter are dropped by an age guard; death dates before birth dates (the graph's typos) ship as NA. The century's defining curve is enforced by test: women's share of participation rises from under 15 percent before 1936 to over 38 percent since 2010. Built with a 15-test acceptance suite.
338,127 rows across 2 joinable files, with 36 columns in total: Athlete, Athlete_ID, Gender, Birth_Date, Death_Date, Birthplace, Country, Citizenship, Sports, Height_CM, Weight_KG, Games_Count, Events_Count, First_Games_Year, Last_Games_Year, Career_Span_Years, Age_At_First_Games, Seasons, Gold, Silver, Bronze, Total_Medals, First_Medal_Year, Last_Medal_Year, Athlete_URL, Athlete, Athlete_ID, Event, Event_ID, Games, Year, Season, Sport, Age_At_Games, Rank, Medal.
It is built by programmatically scraping wikidata.org, last pulled on 2026-09-14. 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.
Join the files on Athlete_ID (participations.Athlete_ID = olympians.Athlete_ID); Events_Count on the athletes file equals that athlete's participation rows exactly.
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.