One row per holder of chess's highest title: all 2,127 FIDE Grandmasters Wikidata knows, from the 27 inaugurated in 1950 to the twelve-year-olds of the 2020s, with the best graph coverage of any topic on the platform (every row carries a title year, 99% a peak Elo, 95% the International Master year too). The columns tell modern chess's whole story. Title years by decade show the inflation: 61 new GMs in the 1950s, 606 in the 2000s. Age_At_GM_Title tracks the prodigy race (Abhimanyu Mishra's 12 leads, and every sub-15 title belongs to 1991 or later). IM_To_GM_Years measures the ladder's hardest step. The rating book distills every dated Elo statement in the graph (Carlsen alone carries 197) into peak Elo with its year, age at peak, the latest rating, and a Rating_Count so analysts know how much curve sits behind each row; Wikidata's monthly imports thin out before 2000, so each article's infobox Peak rating line (FIDE's official figure) fills the historic gap and the larger value wins. Around the numbers: sex (44 women hold the full title), federation, birth and death dates, and the FIDE ID as the outward join key to rating sites. World_Champion flags the classical lineage from Euwe to Gukesh, curated by hand because the graph does not model the championship, and asserted by test. Ages honor stated precision: exact where the graph dates both sides to the month, the year difference otherwise, which can read one year high for a late-year birthday (Carlsen's title is dated only 2004, so his row says 14 though April knew him as 13; this sentence is the disclosure). Built with a 13-test acceptance suite anchored on Carlsen's 2882, Fischer's 1958 title, Mishra's record and the curated champions line.
2,127 rows in one CSV file, with 19 columns in total: Player, Player_ID, Sex, Born, Died, Country, FIDE_ID, IM_Title_Year, GM_Title_Year, Age_At_GM_Title, IM_To_GM_Years, Peak_Elo, Peak_Elo_Year, Age_At_Peak, Latest_Elo, Latest_Elo_Year, Rating_Count, World_Champion, Player_URL.
It is built by programmatically scraping wikidata.org, last pulled on 2026-09-16. 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.