What is a synthetic population?¶
If you want to simulate the real world, you need to populate your simulation with realistic people. The key word is realistic. Modeling real individuals infringes on their privacy. Instead, we constructed a realistic synthetic population that is statistically indistinguishable from the U.S. Census.
Our platform comes preloaded with this realistic synthetic population, which we built using data from a variety of sources:
- 2019 5 year American Community Survey (ACS)
- 2019 Public Use Microdata Survey (PUMS)
- 2019 Census shapefiles for counties, subcounties, tracts, block groups
- 2020 National Center for Education Statistics (NCES)
- 2015 ACS commuting flows
- Facebook Data for Good Initiative Population Density
- 2005 INFOUSA Workplace locations and sizes
Each individual in the population is synthetic, but every census block perfectly matches its real-world equivalent across demographic variables like age, sex, race, income, and occupation. That means you can start building and running powerful models right away without any external data. Plus, if you have your own proprietary data, you can enrich the synthetic population to represent the kind of people you seek to serve with even more precision.