Help build the world's largest open database of politicians.
PoliLoom is a community tool for improving politician information on Wikidata, the free knowledge base behind Wikipedia. We use AI to find missing information about politicians — birth dates, positions held, citizenship, and more.
- We extract — AI reads Wikipedia articles to find politician data that's missing from Wikidata
- You verify — Log in with your Wikipedia account and review extracted information, one politician at a time
- Wikidata improves — Verified data gets submitted to Wikidata, making it available to Wikipedia and thousands of other projects
Journalists investigating corruption, researchers studying democratic backsliding, and civic tech projects tracking representatives all depend on structured political data. But this data is incomplete:
- Many politicians are missing birth dates, positions, or other basic facts
- Information exists in Wikipedia articles but isn't structured in Wikidata
- Manual data entry doesn't scale to 100,000+ political positions worldwide
By spending a few minutes verifying data, you directly improve the world's open knowledge infrastructure.
PoliLoom is open source. To run your own instance:
git clone https://github.com/opensanctions/poliloom.git
cd poliloom
docker compose up -dSee poliloom/README.md for backend setup and poliloom-gui/README.md for frontend setup.
PoliLoom combines large-scale data processing with AI extraction:
- Processes the complete Wikidata dump (100M+ entities)
- Uses OpenAI to extract politician data from Wikipedia articles
- Matches extracted text to Wikidata entities using semantic search
- All extracted data requires human verification before submission
Backend (poliloom/): Python, FastAPI, PostgreSQL with pgvector, SQLAlchemy
Frontend (poliloom-gui/): Next.js, React, TypeScript, Tailwind CSS
PoliLoom is developed by OpenSanctions as part of our mission to make critical data about public figures freely available.