A personal, static, tracker-free dashboard for publicly disclosed congressional stock trades — glowing eyes and all.
Members of Congress must publicly disclose their stock trades under the STOCK Act. Nancy Fire turns those filings into a dashboard you host yourself: no accounts, no third-party trackers profiting from your attention, no backend — just a static page, a JSON file, and a Python script that refreshes it. Filter any House member; the default follows the most famous portfolio in Washington.
And yes: when disclosure momentum runs bullish, the eyes glow and the smoke rolls. Click the mascot for a burst. Serious data, unserious mascot.
| v1 | v2 |
|---|---|
| Tkinter desktop app, blocking scrape at startup | Static web dashboard, deploys to GitHub Pages |
| Scraped a third-party tracker site (2 columns) | Normalizes actual STOCK Act filing data (10 fields incl. amounts, owner, filing links) |
| 5 pip dependencies | Zero dependencies (stdlib fetcher + vanilla JS) |
| Manual refresh | GitHub Action refreshes data weekly, auto-commits |
| Static image tab | Ember-glow mascot, eye glow on bull runs, canvas smoke particles, click interactions |
| Quantity histogram | Momentum gauge, monthly buy/sell volume chart, filterable table, late-filing flags |
Add yours (the mascot image isn't in this rebuild — see step 2 below):
| Dashboard | Bull-run mode | Trades table |
|---|---|---|
 |
 |
 |
git clone https://github.com/moderatedan/nancy_fire.git
cd nancy_fire
# 1. Pull real disclosure data (writes data/trades.json)
python3 fetch_trades.py # default filter: Pelosi
# python3 fetch_trades.py --member "Greene" # any House member
# python3 fetch_trades.py --member "" # everyone (big)
# 2. Add the mascot: keep your original image and rename it
mv 8b8YFRSSshx4Sfy9fk9w--1--0j7p9.png nancy.png # (old filename also works)
# 3. Open it
python3 -m http.server 8080 # → http://localhost:8080Until you run the fetcher, the page shows clearly-labeled fictional demo data (member "J. Sample (DEMO)") so the layout works out of the box — a banner reminds you it's not real.
- Momentum gauge — trailing-90-day estimated buy vs. sell volume from filings. When buying dominates (≥60%), the mascot goes hot: pulsing ember glow, flickering eyes, ambient smoke.
- Monthly activity chart — buys up in green, sells down in red, estimated dollar volume by month (hand-rolled SVG, no chart library).
- Filterable trade table — every filing with ticker, asset, filed amount range, owner as filed, a link to the source PTR where available, and a "filed after" column that flags filings past the 45-day STOCK Act window ⚠.
- Stats row — trade counts, buy/sell split, estimated volume, most-traded ticker.
- Auto-refresh CI —
.github/workflows/update-data.ymlre-runs the fetcher every Monday and commits new filings; GitHub Pages redeploys itself. - Interactions — click the mascot for a smoke burst (with a counter, because why not).
prefers-reduced-motiondisables all animation. - Eye positioning — the glow dots are CSS variables; tune
--eye1-x/yand--eye2-x/yat the top ofindex.htmlto your image in ~30 seconds.
- Source: public STOCK Act periodic transaction reports, via the volunteer-run House Stock Watcher JSON mirror. Public records, reshaped — nothing scraped from commercial trackers.
- It is not real-time. Members have 45 days to file, so every number here lags reality. The dashboard prints this on the page rather than pretending otherwise.
- Amounts are ranges. Filings report brackets like "$1,000,001 – $5,000,000", never exact figures. Volume stats use range midpoints and are labeled estimates.
- Owner matters. Filings often list a spouse or dependent as the transacting owner (famously so for the default member); the owner column preserves exactly what was filed.
- Not financial advice. This is a public-records viewer with a cartoon on it.
The mirror is volunteer-maintained. If fetch_trades.py fails:
- Try again later, or import manually:
python3 fetch_trades.py --csv your_file.csv(accepts the oldtransaction_data.csvformat too). - Alternative sources to adapt: the House Clerk's Financial Disclosure portal (official, PDFs), or a commercial API (e.g., Financial Modeling Prep's House disclosures endpoint) — PRs adding a second source welcome.
winamp.html renders the same data as a classic late-90s media player — playlist of filings, LCD marquee, spectrum analyzer driven by trade size, transport controls — and desktop/nancy_fire_desktop.py (~90 lines of WebKitGTK, no Electron) runs it as a native Linux desktop app with the skin's own draggable title bar. See WINAMP.md.
- Senate support (senate-stock-watcher dataset, same shape)
- Price overlay: trade markers on the ticker's chart
- Multi-member comparison view
- RSS/JSON feed of new filings from the CI run
MIT. Disclosure data is public record.