Data Specifications
Time Coverage
1998–2024
Geographic Coverage
United States (5 large states)
Granularity
Equipment-level (make, model, characteristics, prices)
Update Frequency
Annual
Methodology
The Machinery of Progress dataset is constructed by digitising archival administrative filings from 1998 to 2024 and extracting 50 million capital equipment transactions from five large US states.
We deploy an "agentic AI" measurement approach, where multiple AI agents collaborate to build and validate the data. Specialised agents handle document parsing, entity extraction, classification, and quality checking in a coordinated pipeline.
The final dataset contains the make and model of millions of pieces of equipment — IT equipment, heavy machinery, agricultural tools, vehicles, robotics, CNC machines, and more — along with equipment-level characteristics including time-varying prices. This enables research on technological progress, innovation diffusion, and capital investment patterns at unprecedented granularity.
License & Terms
Commercial license. Public preview dataset available soon. Contact us for licensing.