DatasetPressAbout DatasetPress
Public records deserve better publishing.
DatasetPress is being built to turn difficult, fragmented public records into reliable datasets people can understand and use.
The idea
Enormous amounts of valuable information are technically public but practically hard to use. Records can be spread across agencies, buried in PDFs, formatted differently from year to year, or described with local terminology. Finding the files is only the beginning.
DatasetPress approaches that gap as a publishing problem. The goal is not simply to scrape more rows. It is to create a dependable bridge between the original record and an analysis-ready table.
How projects are chosen
Good candidates have public value, a credible authoritative source, and a meaningful usability problem that careful normalization can solve. Before publication, each project must also have a defensible approach to source rights, privacy, documentation, and ongoing maintenance.
DatasetPress does not claim false precision, government endorsement, customer adoption, or completeness beyond the documented edition coverage.
The first publication
The inaugural 2026 Edition focuses on historical Michigan municipal water and sewer bid prices. It includes 15,814 usable normalized observations, 88 project-source records, 50 agencies represented in usable data, and 133 bidders, delivered with source links and visible QA context.
What comes next
The catalog will grow selectively. Each new dataset should make a hard-to-use public record measurably more usable while preserving the context needed to interpret it responsibly.