Trump Trades Data Downloads Are Up Almost 70% After the Midterms. 20 Trade Reports, 10,000+ Trades and 28,000+ Annual Disclosure Records

Trump Trades Data Downloads Are Up Almost 70% After the Midterms. 20 Trade Reports, 10,000+ Trades and 28,000+ Annual Disclosure Records

Published: November 7, 2026

Downloads of Trump trades data went up almost 70% the past 3 days... after the midterms. Wonder why. The data has all 20 of Trump's trade reports since August 2025, 10,000+ trades, including the latest one posted on 8 Oct. Plus his 2025 annual disclosure, 927 pages converted into 28,000+ records.

For you - Live tool:

tigzig.com/tremor/tools/trump-trades

For your AI agent:

"Go to tigzig.com and download all the Trump data files, including the methodology and validation."

For both - Tigzig downloads hub:

tigzig.com/downloads

Some nuances of the extraction are below. Details in docs.

Some quick points on extraction, details in docs

▸ The trade reports are scanned images with no text inside. Any one reader misreads a digit now and then or slips a row, and nothing flags it. So each page is read three times by separate methods and a value goes in only when two agree. On the latest report one model slipped a row and swapped buys and sells on eight trades, and the vote caught it.

▸ The annual disclosure is an Excel file opened in Google Sheets and saved as PDF. Its own title says so. The numbers are in the file as text, and that makes it far easier. A parser can still push a wrapped line into the wrong column without any error. So two parsers that share no code read it, and every field on 27,000+ rows was compared until the two matched exactly.

▸ The AI agent settles every disagreement against the page image, and I review what it flags before anything goes live.

▸ The annual does not repeat trades already filed on a trade report. In a cross-check of 680 such trades, one showed up in both.

▸ The filings write the same company many ways, such as AMAZON COM INC and AMAZON COM INC COM. To total trades by company each name is mapped to its ticker, about 2,800 names in all, each with a confidence level.


Working on something similar? How I work covers the rates, the availability and what I take on.