Each dashboard is rebuilt on a separated data + spec process: the reconciled data (official PUFs where they exist) streams as columnar parquet and every number is computed in your browser with DuckDB-WASM; the look lives in an editable spec harvested from the original Tableau workbook and rendered by Plotly. Totals pin the published All/Total members, so they match the official dashboards exactly. Click ✎ Edit on any page for point-and-click editing.
EEO-1 (rebuild)
LIVE · PUF 2014–2023 · editable — Employer Information Report — US map, job categories, race/ethnicity, sex by state, 10-year trend. National 2023 total matches Tableau exactly: 77,845,097.
Open →EEO-3 (rebuild)
LIVE · PUF 2012–2022 · editable — Local Union Report — members, applicants and referrals by state, type, race & sex.
Open →EEO-4 (rebuild)
LIVE · PUF 2005–2023 · editable — State & Local Government — full-time employees by state, job category, race & sex.
Open →EEO-5 (rebuild)
LIVE · PUF · editable — Elementary–Secondary Staff — staff by state, role, race & sex.
Open →EEO-1 Pay Data (rebuild)
LIVE · Component 2, 2017–2018 · editable — Salary bands by state, industry, job category, race & sex from the Component 2 pay collection.
Open →Charge Statistics (rebuild)
LIVE · FY1992–2025 · editable — Charge receipts by state and year, merit resolutions by charge basis.
Open →Litigation (rebuild)
LIVE · FY1997–2025 · editable — Suits filed by year and the selected-resolutions case list.
Open →Data Tables (rebuild)
LIVE · FY1992–2025 · editable — Charge, litigation & ADA data tables — trends, statute breakdowns, and the full browsable tables.
Open →EEO-1-Convert-0 (converted)
62 charts · rebuild method · editable — made with the converter below. Lives until the next redeploy; download the bundle and commit it to keep it permanently.
Open → Bundle ↓Convert a dashboard with the rebuild method
Drop a dashboard file and/or its data. The converter separates data from spec: data becomes parquet streams, the look becomes an editable spec, and the result opens here with the same in-browser engine and edit mode as the dashboards above. A Tableau workbook provides both look and data; a Power BI (.pbix) report's charts are rebuilt from its report layout by an AI agent (and R Shiny from its data) — add the data as separate files (a .pbix may carry its own CSV exports) since Power BI hides its data in a format we cannot read. The rebuild method never invents data it cannot read.
