UnPAC - Claude Vibeathon Spring 26 Beginner Track Winner

UnPAC lets you enter any U.S. address and instantly see every representative who serves that address — what they vote for, who funds them, and whether their votes align with their donors.

What it does

1. Enter your address — the app resolves your federal, state, and local representatives via the Google Civic Information API.
2. See your reps — each representative shows their party, office, photo, and a quick-glance influence score.
3. Drill into a rep — click any representative to open a detailed panel with:
   • Voting record — recent bills sponsored, co-sponsored, and voted on, categorized across 8 issue areas (Healthcare, Energy, Defense, Economy, Education, Immigration, Environment, and more)
   • Who funds them — campaign finance data from the FEC broken down by industry sector, with a bubble chart and ranked contributor list
   • Influence scores — two data-driven scores (see below)
   • AI conflict analysis — on demand, Claude (Haiku) reads the voting record against donor industries and surfaces specific conflicts of interest

Influence scores

Legislative Influence Score (0–100)

Measures how financially captured a legislator is by industries with active regulatory exposure (finance, healthcare, energy, defense, etc.).

Uses an HHI-style concentration formula: a rep whose donations are dominated by a single industry scores higher than one with the same total spread across many sectors. A power curve pushes mid-range scores outward so the distribution is readable at a glance.

• 0 — donations distributed across many non-regulatory industries
• 100 — heavily concentrated in one high-influence sector

Voting Alignment Score (0–100)

Measures whether a rep's floor votes and bill sponsorships align with the interests of their donor industries.

Voting actions are weighted by signal strength:

Action	Weight
Floor vote Yes	+1.0
Floor vote No	−1.0
Sponsored	+0.4
Co-sponsored	+0.2
Not Voting / Present	0

Donor sectors are mapped to bill categories (e.g. "pharma" → Healthcare, "oil/gas" → Energy). The score is a donor-weighted average of alignment across matched categories.

• 0 — consistently votes against donor industry interests
• 50 — neutral or no clear pattern
• 100 — consistently votes in favor of donor industry interests

AI conflict analysis

Clicking "Analyze Conflicts of Interest" sends the rep's top 30 weighted legislative actions and top 8 donor sectors to Claude (Haiku). Claude returns a nonpartisan JSON report classifying the overall tension level (High / Medium / Low / None) and listing specific industry-by-industry conflicts with explanations tied to named bills.

Data sources

Data	Source
Representative lookup	Google Civic Information API
Federal legislators & bills	Congress.gov API
State legislators	OpenStates API
Campaign finance	FEC API
AI analysis	Anthropic API (Claude Haiku)

Running locally

Copy .env.local.example to .env.local and fill in your API keys:

CONGRESS_API_KEY=
FEC_API_KEY=
GOOGLE_CIVIC_API_KEY=
OPENSTATES_API_KEY=
ANTHROPIC_API_KEY=

Then:

npm install
npm run dev

Open http://localhost:3000.

Tech stack

• Next.js (App Router) — frontend and API routes
• Tailwind CSS — styling
• Anthropic SDK — conflict-of-interest analysis
• Data fetched server-side in API routes to keep keys off the client