An independent, bottom-up ledger built from the City of Austin’s own published checkbook. We start from the checks the city actually cut, sort each one into the city’s own fund types, and show who received the money and whether it traces back to a contract. Every figure links back to a city source so you can verify it yourself.
We deliberately show a slice, and we’re explicit about it:
8c6z-qnmj →. Numbers move as the city publishes new data.Every check is tagged on two separate axes, because “where the money came from” and “what it bought” are different questions:
Axis 1 — Fund type & program (“where it came from / where it went”). We map each check’s fund name to the City of Austin’s own governmental fund taxonomy (below), then to a program within that fund (e.g. Live Music Fund, Austin Water). This is normalization — decoding the city’s cryptic fund codes by their stable naming conventions — not a hand-maintained list of projects, so new funds that follow the same naming fall into the right bucket automatically.
Axis 2 — Purpose (“what it bought”). The checkbook’s object/commodity code (e.g. Construction Services, Loan Administration). This is the city’s internal bookkeeping label and is shown as secondary detail — it describes the transaction type, not who benefited.
The City’s own governmental fund types. “Approx scale” is a rounded FY2025-26 reference for context ($6.3B all-funds Approved Budget →), not a precise per-fund appropriation.
Order is load-bearing (first match wins) so real collisions resolve correctly — e.g. a debt fund carries the utility name it repays, so debt is matched before enterprise. The exact rules live in lib/funding.ts (classifyFund and fundSubGroup).
Within each fund we split the money two ways: traceable (the check names a contract or a delivery order under a master agreement) vs. direct (no source document referenced at all). A large direct payment is the core oversight signal — real money we can’t tie back to a competed contract — so we make that gap visible rather than hiding it.
Under each program we roll up the actual recipients (vendors) and how much each received, biggest first. Each links to a per-vendor page showing every check to them in the snapshot, plus deep links to verify at the city’s source.
Of the $4,383,235,262 in scope, 100% maps to a named city fund type. Anything we can’t confidently map goes into an explicit “To be categorized” watch-list — it is never hidden or silently dropped. We’d rather show a small unmapped pile than guess.
Police — $690,345 (needs a human look at its funding source)Aviation — $354,291 (needs a human look at its funding source)Before trusting the numbers we trace real data through the classifier, not just review the code. scripts/audit-categorization.mjs runs every distinct fund name through the same rules and prints fund → bucket → program with dollar totals, flagging anything unmapped. We re-run it whenever the classifier or the data changes. Decision records and the latest audit findings are kept in the project’s docs/ folder.
City of Austin checkbook dataset (8c6z-qnmj) →
Austin Finance Online — solicitations & contracts →
Austin Legistar — legislation & contract awards →
City of Austin FY2025-26 Approved Budget ($6.3B all funds) →
This is a filtered, dated snapshot of large vendor payments — not the full budget, not real-time, and not an official city product. Fund classification is best-effort normalization of the city’s published fund names; the “traceable” link reflects what each check references in the checkbook, which the city’s own data does not always populate. When in doubt, verify at the sources above.