# Start here

You are looking at the evidence archive behind a published *Public Source* story on Pittsburgh's inclusionary-zoning fight. Three ways in, depending on why you came:

- **To read the piece:** open `site/index.html` in a browser. It carries the finding, one chart, and links to everything else.
- **To check the numbers:** run `python 03-harness/check.py`. It needs nothing but Python.
- **To use the data:** open `spreadsheet/Pittsburgh-IZ-Record.xlsx`, or read the CSVs in `02-data/`.

## The one-paragraph version

Pittsburgh requires new developments of 20 or more units to make 10% of them affordable. The mayor wants to extend that rule from four neighborhoods to ninety. At a September 2025 hearing, seventy people testified and both sides claimed the data supported them, though the city has never published how many affordable units the rule has actually produced. This archive answers that question the only honest way available: a project-by-project ledger that separates what is built (35 units) from what is under construction (16), what is merely proposed (100-plus, claimed), and what was redrawn to 19 units to avoid the rule entirely (3 projects). Each figure traces to an archived, hashed source.

## Orientation

- The dataset is the point. `02-data/iz_projects.csv` is the compilation the city never made.
- The dispute is preserved, not resolved. `contested_claims.csv` and `FACT-CHECK.md` state what each source supports and where it stops.
- The trust is mechanical, not personal. `03-harness/` lets a reader re-derive every number without trusting the reporter.

Next: `README.md` for the reader contract, `METHODOLOGY.md` for how the ledger was built.