Methodology
How the machine reads disclosed congressional trading data.
Capitol Alpha Machine turns public disclosure records into cinematic data scenes: disclosed trading volume, member exposure, market movement, and timing proximity to public-law events.
Sources
The primary trading source is Capitol Trades public disclosure data. Returns Spectrum combines the disclosure bundle with end-of-day market price signals. Timing Lattice also uses Congress.gov public-law listings for the 118th and 119th Congresses.
Value ranges and volume
Congressional financial disclosures use reported value ranges, not exact order sizes. The project uses representative disclosed values for visual weight and ranking. Capital Flow separates scale from direction: gross disclosed volume sums buys and sells, while net flow treats buys as positive and sells as negative.
| Input | How it is used | Caveat |
|---|---|---|
| Disclosed value ranges | Representative values drive visual weight and summary tables. | Not exact transaction accounting. |
| Gross disclosed volume | Shows total public trading scale across buys and sells. | Not profit, performance, or intent. |
| Net cumulative flow | Buys are positive and sells are negative. | Directional visualization only. |
disclosed value; gross volume sums buys and sells; net flow uses buys positive and sells negative
Timing and sector heuristics
Committee exposure and Timing Lattice are triage heuristics. They are designed to help readers inspect patterns, not to make claims about motive or inside information.
- Committee-sector exposure compares a member's committee context with sectors represented in disclosed trades.
- Timing Lattice links public-law events to trades by date proximity and sector proximity.
- Returns Spectrum compares disclosed trades with trailing market movement for visualization context only.
- Reporting dates can lag transaction dates, and some disclosures include spouse, dependent, or joint transactions.
Caveats
Congressional financial disclosures are incomplete as a market dataset. Some transaction types are excluded from buy/sell flow, reported values are ranges, and disclosure timing can vary. The current bundle skipped exchange and received transaction types for flow math because they are not clean buy/sell signals.
| Known limitation | Current handling |
|---|---|
| Value ranges | Use representative values for visualization. |
| Exchange / received rows | Excluded from buy/sell flow when not directionally comparable. |
| Anomalous dates | 2 rows skipped in the current flow bundle. |
| Timing links | Presented as proximity heuristics only. |