Hi — I've been working on an open-source project called "Know Before You Vote" that cross-references Epstein document entities against current election candidates. The idea is that a voter enters their zip code and sees which candidates on their ballot appear in the Epstein documents, with links to source records.
Your project's categorized entity data from 19,154 FOIA documents is one of the 4 community databases the pipeline is designed to ingest. I've written an ingestion module that parses your focused_entities.json and extracts categorized entities (political, legal, business, etc.) for cross-referencing. It's already been tested against your live data — 47 entities extracted successfully.
I'm unable to continue development, but the repo has working code, 55 passing tests, and a detailed development plan. If you or anyone in your community is interested in finishing it, everything is public domain:
https://github.com/wjcornelius/know-before-you-vote
The key design feature is multi-source corroboration — a connection only appears if 2+ independent databases (yours + at least one other) agree. This prevents false accusations from database noise or NER errors.
No pressure at all — just wanted to put it on your radar in case it's useful.
Hi — I've been working on an open-source project called "Know Before You Vote" that cross-references Epstein document entities against current election candidates. The idea is that a voter enters their zip code and sees which candidates on their ballot appear in the Epstein documents, with links to source records.
Your project's categorized entity data from 19,154 FOIA documents is one of the 4 community databases the pipeline is designed to ingest. I've written an ingestion module that parses your focused_entities.json and extracts categorized entities (political, legal, business, etc.) for cross-referencing. It's already been tested against your live data — 47 entities extracted successfully.
I'm unable to continue development, but the repo has working code, 55 passing tests, and a detailed development plan. If you or anyone in your community is interested in finishing it, everything is public domain:
https://github.com/wjcornelius/know-before-you-vote
The key design feature is multi-source corroboration — a connection only appears if 2+ independent databases (yours + at least one other) agree. This prevents false accusations from database noise or NER errors.
No pressure at all — just wanted to put it on your radar in case it's useful.