Data Collection

We collect public records from government websites across Rockland County, including town and county budget documents, legislature agendas and minutes, planning and zoning board records, campaign finance filings from the New York Board of Elections, school district data from NYSED, and vendor contract information.

Records are collected using automated software that checks government websites on a regular schedule. When source documents are PDFs, text is extracted and structured programmatically. When source pages are HTML, relevant data is parsed directly.

Database

Collected records are organized into a structured database. The current database contains over 47,000 financial records spanning fiscal years 2021–2026, more than 8,200 campaign finance entries, approximately 3,000 meeting documents from all five towns and the county legislature, and school district enrollment and spending data from NYSED for all eight Rockland County districts.

Cross-Referencing

The database supports queries that span multiple data sources. Examples include comparing per-pupil spending across all eight school districts, identifying entities that appear in both campaign finance records and vendor contract records, flagging budget line items where actual spending deviates significantly from appropriation, and tracking spending trends across fiscal years with inflation adjustment.

Analysis and Publication

Cross-reference findings, anomalies, and patterns are reviewed editorially before publication. Evidence Briefs include explicit confidence levels, source citations for every factual claim, statements of what evidence would change the conclusion, and disclosure of data gaps.

We do not publish findings based solely on automated outputs. All analysis passes through editorial judgment before it reaches the site.

Limitations

Our database is not complete. We are continuously expanding coverage and improving data quality. Specific limitations include: not all historical records have been digitized and ingested, some municipalities publish records in formats that are difficult to parse consistently, cross-referencing depends on entity name matching which can produce both false positives and missed connections, and records are only as current as the most recent collection cycle.

When we know our data is incomplete on a topic, we say so in the published analysis.

Source List