Skip to content
ctzn.pub
A verified-data publishing platform

Verified public data — as essays for people and surfaces agents can cite.

ctzn.pub is a publishing platform for researchers, students, and journalists: rigorous public surveys — GSS, ANES, CES, CDC, Census and more — computed correctly and published as interactive essays, with a free public layer and an institutional library for researchers. Every figure survey-weighted, every claim sourced.

The Data Pipeline

From raw public survey data to a published essay. Four stages, the same for every article.

Public Sources

Public survey and administrative microdata — GSS, ANES, CES, YRBS, MTF, SCF, CDC PLACES, Census/ACS and more — pulled from the canonical source and pinned to a versioned extract.

Harmonization

Content Engine

Machine Surfaces

What Makes This Different

The four things we won't compromise on.

Academically Rigorous

Every figure is survey-weighted and every claim traces back to its exact source — no vibes, no invented numbers.

Pinned to a Snapshot

Every figure is pinned to a versioned data extract — not a live query that can drift under you. We refresh the pin when a source publishes a new wave, and every article states its vintage.

Agent-Ready Surfaces

Markdown mirrors, a machine-readable claims feed, and JSON-LD — the surfaces AI engines cite today.

Built to Read

Interactive charts and maps designed for long-form reading — built for researchers and journalists first, machine-readable second.

Who runs this

ctzn.pub is built and run by a small team working directly with the survey methodology — the same people who compute the figures, write the essays, and review every published number.

Contact

Questions about a figure, a correction, a licensing question, or interested in the institutional library — reach us directly.

info@ctzn.pub
For agents & builders

Machine surfaces

What machines can read today: Markdown mirrors of every published article, a machine-readable claims feed, and structured data on every article page. No key required — these are public, crawlable surfaces.

Markdown mirrors

Every published article is also served as clean, extraction-friendly Markdown — prose intact, each chart replaced by its sourced takeaway.

GET /api/md/<username>/<slug>

Claims feed

A machine-readable feed of sourced findings from published articles — the numbers, their source surveys, and the article they came from.

GET /claims.json

Structured data

Published articles carry Schema.org JSON-LD (Article + Dataset), so search and answer engines can attribute every figure.

application/ld+json on article pages

Open to AI crawlers

robots.txt explicitly welcomes GPTBot, ClaudeBot, PerplexityBot and peers across all public content.

GET /robots.txt
claims.json — live example
{
  "generated": "2026-07-03T22:14:36.012Z",
  "count": 2,
  "claims": [
    {
      "article": "The widening mental health gap among US teens",
      "url": "https://uma.ctzn.pub/the-widening-mental-health-gap-among-us-teens-wyl28",
      "title": "Poor Mental Health Among US High School Students",
      "source": "YRBS",
      "takeaway": "Female: 40.8 (2021) → 38.8 (2023), -2.0; Male: 18.1 (2021) → 18.8 (2023), +0.8."
    }
  ]
}

On the roadmap

A full public REST API and MCP server over the article and claims catalog is in development and not yet available. Nothing on this page requires a key or a plan — when a metered API ships, it will be documented here first.