This invention is a computational system that generates high-resolution synthetic populations representing individuals, households, and their socioeconomic characteristics for the entire United States. It provides researchers and planners with a highly accurate, privacy-preserving dataset that matches real-world marginal distributions at the Census block group level.
Description
The technology utilizes an Iterative Proportional Updating (IPU) algorithm to harmonize data from the Decennial Census, American Community Survey (ACS), and Public Use Microdata Sample (PUMS). By using PUMS data as a "seed" and Census marginals as "targets," the system iteratively adjusts weights to ensure that both person-level (e.g., age, race, sex) and household-level (e.g., income, size) characteristics are jointly matched with high precision.
This approach leverages LandScan nighttime data to assign precise latitude and longitude coordinates to these synthetic households. The method employs bidirectional score-based integerization to convert fractional weights into discrete individuals and households while maintaining the statistical integrity of the original Census data. This results in a granular, micro-level population dataset that reflects the complex compositions of real-world communities.
Applications
- Public Health & Epidemiology: Simulating disease spread or evaluating the impact of healthcare interventions on specific demographic groups.
- Urban Planning & Transportation: Modeling commuting patterns, school district requirements, and infrastructure needs based on localized population growth.
- Market Research & Consumer Analytics: Enabling companies to perform hyper-local market segmentation and demand forecasting without purchasing expensive proprietary data.
- Disaster Response & Emergency Management: Estimating the number and types of households in high-risk zones for better evacuation and resource allocation planning.
- Policy Impact Simulation: Testing the potential effects of economic or social policies on diverse household types before implementation.
Advantages
- Great Accuracy: Achieves highly accurate alignment with official Census targets, with total population counts generally within narrow margins of the reported figures.
- Multi-Level Matching: Simultaneously matches both individual and household attributes, ensuring realistic family and living structures.
- High Spatial Resolution: Provides geographic precision down to the Census block group level with specific coordinate assignments.
- Privacy-Preserving: Generates realistic synthetic individuals, allowing for detailed modeling without compromising the actual identities of residents.
- National Scalability: Designed to process and generate data for the entire United States, making it suitable for large-scale federal or regional projects.
Invention Readiness
The technology has successfully completed the data harmonization, IPU algorithm execution, and household coordinate assignment phases for the entire U.S.. Validation data shows highly successful matching across various spatial units, including block groups, counties, and states. Ongoing development is focused on creating group quarters populations, assigning workplaces to adults, and assigning schools to students to create a fully functional social fabric model.
IP Status
Copyright
