July 2026 register contains July 2025 dates
The city page labels the workbook as July 2026, while every transaction inside falls between July 1 and July 31, 2025. The file is quarantined from totals.
Investigation workspace
Find the pattern. Preserve the context.
This queue surfaces source conflicts, missing detail, concentration, recurring amounts, and project sequences. Each signal is a question to inspect—not a verdict.
What makes a signal useful?It has a reproducible rule, a visible denominator, a path to the source, and language that does not overstate what the data proves.
89.4% of FY2026 payments have no memo
4,409 of 4,934 published transactions do not explain what was purchased. Payee and amount remain available, but purpose-level analysis is limited.
Five payees account for 47.3% of published payments
The concentration is driven by debt, retirement, payroll tax, and a large construction contractor. This is a review signal, not an allegation.
Two Goodwin / Conrads payments total $3.54M in June
Capital Excavation Company received payment applications #15 and #16 twelve days apart. The shared project reference makes this a traceable project sequence.
Rules running now
A review engine people can audit
Every rule will publish its logic, inputs, exceptions, and last-run time.
Source-date agreementCompare document label, coverage period, and dates inside each workbook.
Missing description rateCount blank payment memos by month, method, and payee.
Payee concentrationMeasure top-recipient shares with government and benefits context.
Repeated project referencesGroup explicit memo tokens such as payment-application numbers.
Fiscal year-end movementCompare late-year payment patterns against prior periods.
Threshold clusteringReview amounts near policy thresholds only after extracting the policy.
Signals are generated only from published records and disclosed rules. Nothing on this page alleges waste, fraud, illegality, or misconduct.