CDB90: A tidy, open-data version of the CAA Database of Battles (1990)

A tidier version of CDB90, the US Army Concepts Analysis Agency Database of Battles (Version 1990, Helmbold 1993): 660 battles fought between 1600 and 1973, with their strengths and losses, outcomes, dates, durations, and tactical and environmental descriptors. This deposit holds the data only, in three forms: flat tables for analysis as CSV and Parquet (battles: one row per battle; sides: one row per battle side; narratives: the narrative, significance and sources of 610 battles); a relational SQLite database (cdb90.sqlite3) with the tidy tables, code tables, links and coordinates; and the same relational tables as CSV (relational-csv.zip). A Frictionless Data Package (datapackage.json) describes the flat CSV tables. 658 of 660 battles are linked to one or more Wikidata items and Wikipedia articles, and 655 of 660 have coordinates from Wikidata. The code that builds the data is on GitHub (https://github.com/jrnold/CDB90, MIT). The original data are public domain; the derived data are CC0.