DJIN model of aging synthetic dataset

The DJIN model of aging was trained on the English Longitudinal Study of Aging (ELSA). Here we have used the model to generate a large synthetic population of 9 million individuals. There are 3 million individuals for each baseline age of 65, 75, and 85 years simulated for 20 years. For each individual, we supply a health trajectory with 29 tracked health variables with mortality. Demographic and background health variables have been sampled based on the ELSA population demographics.

Each Data_part includes 1.8 million individuals. The file_description.txt file describes the files, and health_columns.csv and background_columns.csv indicate the columns of the files.

The ELSA dataset itself can be accessed at https://www.elsa-project.ac.uk/accessing-elsa-data.

Code for the model is available at https://github.com/Spencerfar/djin-aging.