Congressional Tech

Open-source tools for a more transparent Congress.

A Civic Tech DC initiative, in partnership with the American Governance Institute, building small, practical tools that make congressional data and operations easier to find, track, and understand — for congressional staff and the public alike. Every proposal here started as a real request from people working in and around Congress.

🔗 Live site: https://civictechdc.github.io/congressional-tech/ — browse the project catalog, read the proposals, and explore the committee YouTube coverage dashboard.

What's in here

This is a Turborepo monorepo using npm workspaces, in two layers.

apps/ — things you run or deploy

App	What it is
site	The public website (Astro). Project catalog, proposals, and interactive dashboards. Deployed to GitHub Pages.
committee_youtube	CLI tools that scan congressional committee YouTube channels and flag videos missing the official Event ID that links them back to Congress.gov.
inflation_gsheets	Pulls inflation data (BLS CPI + FRED alternative measures) into CSVs that feed a Google Sheets inflation calculator.

packages/ — shared Python libraries the apps build on

Package	What it provides
congress_shared	Auth, config, and reference data shared across the tools.
congress_api	Production meeting mirror, GovInfo and committee-site capture, inventories and transcripts. Start with congress-meetings; TinyDB explore commands remain optional.
youtube_api	Fetch and analyze committee YouTube data. CLIs: youtube-fetch, youtube-analyze.
committee_meeting	Committee meeting metadata helpers.
corpus_recovery	Checks recovered committee documents and adds them to the raw archive. Commands are in scripts/corpus_recovery.
viewer_publication	Publishes the public filename viewer without rewriting archive objects. Commands are in scripts/publication.
news_lake	Political news lake: record contract, Iceberg storage boundary, and offline adapter/identity/analysis cores from the design packet (proposed; see docs/adr/news-lake-contract.md).

The project catalog

The work is grouped into five themes. The site renders 19 proposals (each with a problem and a proposed solution) plus the tools already shipped. The source of truth is plain markdown you can read right here on GitHub: apps/site/src/content/.

#	Theme
I	Enhancing Congressional Data Accessibility & Usability
II	Workflow Efficiency Tools for Congressional Operations
III	Leveraging AI for Content Enhancement & Analysis
IV	Predictive & Analytical Tools for Legislative Insight
V	Basic Utility Tools

Two proposals are working tools today: the Committee YouTube coverage dashboard (I.2) and the Google Sheets inflation calculator (V.1).

Sub-projects

Two sister tools built by AgoraDMV, forked into this org and kept in sync with upstream daily:

• DeltaTrack — structurally diffs U.S. bill versions from GPO govinfo (added / removed / modified / moved sections), with an account-level old→new money table for appropriations. Local and zero-dependency, built for Hill staffers.
• BillTrax — a public legislative-intelligence platform: ingests Congress.gov data, structures bill text, tracks funding changes across versions, and adds AI summaries and topic tags.

Getting started

Prerequisites: Node 20+, Python 3.12.13, uv, and Poppler for PDF validation.

The website

cd apps/site
npm install
npm run dev      # serves http://localhost:4321/congressional-tech/
npm run build    # static output in dist/

Content lives as markdown under src/content/ (themes, proposals, projects) — add or edit a .md file and the site picks it up.

The Python tools

One locked environment installs the workspace Python packages:

uv sync --locked
source .venv/bin/activate
python scripts/dev.py doctor
python scripts/dev.py prepare-baseline
python scripts/dev.py test production-offline
python scripts/dev.py test active-experiment-regression

Other commands in scripts/ publish the filename viewer, add recovered documents to the archive, query the Library of Congress Web Harvest, and measure local rebuilds. Those jobs are separate from the CLIs below.

Install Poppler with brew install poppler on macOS or sudo apt-get install poppler-utils procps on Ubuntu. See the development guide for test lanes, native checks and frozen research boundaries. The CLIs are then on your PATH:

youtube-fetch --help
youtube-analyze --help
congress-meetings --help
congress-committees --help
meeting-inventory --help

The weekly production sequence and saved-data formats are documented in congress-api and its compatibility reference. Help does not require credentials. The unused Congress exploration tools were retired after preserving their useful acquisition behavior.

Inflation data:

# BLS CPI index grid -> src/data/historical-cpi.csv
python apps/inflation_gsheets/src/fetch-cpi-convert-csv.py

# BLS + FRED alternative measures (+ optional Adobe/PriceStats) as YoY% -> src/data/inflation-sources.csv
python apps/inflation_gsheets/src/fetch-inflation-sources.py --aggregate

Dev container

The repo ships a dev container that installs the Python graph and Node dependencies for you. In VS Code, choose Reopen in Container.

Automation

Workflow	Cadence	What it does
capture-raw-sources.yml	daily, manual, or validation on code changes	Retains source evidence and rebuilds document tables; exposes a run-bound heartbeat.
deploy-pages.yml	on push to main	Builds apps/site and deploys it to GitHub Pages.
bls-cpi-update.yml	monthly	Refreshes the CPI baseline and the multi-source inflation table.
update-data.yml	weekly	Refreshes YouTube, GovInfo, Congress.gov, committee-source and inventory data.
publish-explorer.yml	after collection completes, source changes, or manual dispatch	Publishes and verifies Explorer data from retained inputs before deployment.

Inspect an existing capture or rebuild run without changing it:

.venv/bin/python scripts/publication/pipeline_status.py --run RUN_ID

The pipeline status guide explains identity checks, expired heartbeats, saved evidence and the separate publication-verification step.

Contributing

Come build with us. Pick a proposal from the catalog, open an issue, or say hello in the Civic Tech DC community. Newcomers welcome.

License

See LICENSE.