1. Data Sources
- Based on work from MIT's Project Iceberg.
- Workforce Data: US Census Bureau, American Community Survey 5-Year Data (2023).
- AI Risk Scores: Based on Eloundou et al. (2023) "GPTs are GPTs", mapping automation feasibility to O*NET job codes.
2. Project Goal: The "Digital Twin"
Popular narratives focus on AI disrupting tech hubs (Silicon Valley). This tool creates a "Digital Twin" of local economies
to reveal that the largest volume of exposure is actually in "submerged" administrative, logistical, and middle-management roles found in every zip code.
3. Privacy & Normalization (The "Small Town" Problem)
The US Census Bureau suppresses specific salary data for small zip codes to protect privacy (e.g., if a town only has 3 Nuclear Engineers, publishing their average salary would reveal their personal income).
To solve this, this tool uses National Average Salaries as a baseline, but applies a Local Cost-of-Living Multiplier based on the Zip Code's Median Household Income. This allows us to visualize economic vulnerability without violating federal privacy standards.
4. Why 2023 Data?
We use the ACS 5-Year Estimate (2019-2023) because it is the most recent dataset statistically significant for small populations (under 65,000). While 2024 data exists for large cities, using it would break the tool for rural areas and small towns.