Transparent Politics

Completed
React
ChromaDB
FastAPI
RAG
Python
Civic engagement platform for exploring Texas Congressional districts and legislative tracking via a RAG-powered search pipeline over congressional bill data.
Author

Aakriti Dhakal

Published

January 10, 2026

A data platform to inform citizens about incumbent representatives. Enables users to explore 38 Texas Congressional districts via interactive maps, candidate profiles, and legislative tracking across 100+ policy-categorized bills.

RAG pipeline

Built an ETL pipeline to embed and index congressional bill data in ChromaDB:

  1. Scrapes bill text and metadata from Congress.gov API
  2. Chunks documents by section with overlapping context windows
  3. Embeds using text-embedding-ada-002 and stores in ChromaDB with metadata filters (district, party, topic)
  4. FastAPI endpoint accepts natural language queries, retrieves top-k chunks, passes to GPT-4 for synthesis

Filtering

Metadata filtering lets users scope searches to specific districts or policy areas. A query like “education bills in District 21” retrieves only district-21-tagged education documents before embedding similarity is computed.