100-m resolution age-stratified population grid data for China in 2020

A 100m age-stratified gridded population dataset for the 2020 China Seventh Population Census was generated using Random Forests in combination with geospatial big data. The dataset incorporated county-level and township-level data from the 2020 census, along with six covariates as input features. The Random Forest model was trained and estimated using five-fold cross-validation to generate the age-stratified gridded population data. The estimated population dataset outperforms the existing WorldPop product in accuracy, with R² values for the age groups 0–14 years, 15–59 years, ≥60 years, and ≥65 years improving from 0.48, 0.57, 0.52, and 0.52 to 0.95, 0.95, 0.88, and 0.86, respectively.