A global dataset of immigration flow composition by age, sex and educational attainment

Estimates and R codes of immigration flow proportions by age, sex and education of 199 countries between 1990-1995 to 2015-2020 based on methods presented in Yildiz and Abel (2026). Version v2: Updated R scripts to extend the range of Random Forest models considered in the main SL algorithm and to include a more diverse SL algorithm, and revised output table. Immigration composition by SL provides the results by the Super Learner model with six columns for each of the data dimensions: iso3c: Destination ISO three letter country code period: Five-year periods age: Five-year ge groups sex education: Educational attainment prop: Proportion of immigration flow.