Civic Intelligence — Ontology Crosswalk
Schema: schemas/civic-intelligence/v1.0.md
Target standards: DCAT-US v3.0 (primary), PROV-DM (provenance), schema.org/GovernmentService (secondary)
Authority: U.S. Department of Commerce (DCAT-US is federally mandated by the OPEN Government Data Act for 2026)
Status: FAIR-aligned, Level 3 DCAT-US conformance (per §5)
Dataset SPDX license: CC-BY-4.0 (research tier) · proprietary commercial license for paid tiers (see LICENSE.md)
1. Why DCAT-US v3.0
The OPEN Government Data Act (Pub. L. 115-435) requires federal data to be published in DCAT-US as of FY2026. Any federal, state, or enterprise buyer evaluating Codex will need to confirm DCAT-US conformance before ingestion. Codex publishes Civic Intelligence with both a structured Parquet dump and a DCAT-US v3.0 catalog.jsonld sidecar so the dataset is discoverable by data.gov-compatible crawlers.
2. Dataset-level crosswalk (DCAT-US v3.0 dcat:Dataset)
| DCAT-US field | Codex source | Value |
|---|---|---|
dct:identifier |
fixed | urn:axiom:codex:dataset:civic-intelligence:v1.0.0 |
dct:title |
fixed | "Axiom Civic Intelligence Dataset" |
dct:description |
fixed | "Normalized council proceedings, permits, zoning hearings, and civic filings across U.S. metros, with LLM-extracted entities, blockers, and predictive scores." |
dct:publisher |
fixed | Axiomancer Labs LLC |
dct:issued |
codex_releases.issued_at |
ISO-8601 date of initial release |
dct:modified |
codex_releases.latest_modified |
Latest snapshot publish date |
dcat:keyword |
fixed | ["civic", "zoning", "council", "permits", "government", "transparency", "planning"] |
dcat:theme |
fixed | ["http://publications.europa.eu/resource/authority/data-theme/GOVE"] |
dct:spatial |
derived | GeoJSON union bbox of all records |
dct:temporal |
derived | {start: MIN(occurred_at), end: MAX(occurred_at)} |
dct:accessRights |
fixed | http://publications.europa.eu/resource/authority/access-right/PUBLIC (research tier), RESTRICTED (commercial) |
dct:license |
fixed | https://creativecommons.org/licenses/by/4.0/ (research), Codex Commercial (paid) |
dcat:distribution |
per-format | Parquet, Markdown-KV, CSV (see §3) |
3. Distribution crosswalk (dcat:Distribution)
Each published format gets a dcat:Distribution entry in the catalog sidecar.
| Codex format | dcat:mediaType |
dct:format |
|---|---|---|
| Parquet | application/vnd.apache.parquet |
PARQUET |
Markdown-KV llm_text |
text/markdown |
MARKDOWN_KV (custom; documented in NORMALIZATION_STANDARD.md §5) |
| CSV | text/csv |
CSV |
| JSON-LD manifest | application/ld+json |
JSONLD |
4. Field-level crosswalk
Every Civic Intelligence field maps to a target-standard equivalent where one exists. Fields with no direct equivalent are flagged as Codex extensions and documented in §6.
4.1 Identity & provenance → DCAT-US + PROV-DM
| Codex field | Target | Target field |
|---|---|---|
record_id |
DCAT-US | dct:identifier (record-scope) |
chunk_id |
Codex extension | — |
source_uri |
DCAT-US | dcat:accessURL |
source_system |
PROV-DM | prov:wasGeneratedBy (references prov:Activity for ingest) |
schema_version |
PROV-DM | prov:used (references the schema entity) |
normalization_version |
PROV-DM | prov:wasInfluencedBy (references the pipeline entity) |
ingested_at |
PROV-DM | prov:generatedAtTime |
modified_at |
DCAT-US | dct:modified |
occurred_at |
PROV-DM | prov:startedAtTime of the real-world activity |
published_at |
DCAT-US | dct:issued |
acl_tier |
DCAT-US | dct:accessRights |
provenance[] |
PROV-DM | prov:Activity chain |
4.2 Civic semantics → schema.org + DCAT-US
| Codex field | Target | Target field |
|---|---|---|
document_type |
schema.org | GovernmentService.serviceType (vocab aligned) |
source |
DCAT-US | dct:source |
committee_name |
schema.org | GovernmentOrganization.name |
jurisdiction_slug |
DCAT-US | dct:spatial (jurisdiction IRI) |
h3_index |
Codex extension + OGC | See §4.5 |
occurred_at |
schema.org | Event.startDate |
summary |
schema.org | Event.description |
raw_text |
schema.org | Event.text |
sentiment_polarity |
Codex extension | — |
4.3 Entities → schema.org
entities_extracted[] maps to an array of schema.org/Person or schema.org/Organization linked via schema.org/CreativeWork.mentions:
{
"@context": "https://schema.org",
"@type": "Event",
"mentions": [
{
"@type": "Person",
"name": "Kenyatta Johnson",
"jobTitle": "Council Member",
"identifier": "urn:aprs:entity:person:k-johnson-phila",
"additionalType": "civic.role.councilmember",
"interactionStatistic": {
"@type": "InteractionCounter",
"interactionType": "https://axiomcodex.io/sentiment",
"userInteractionCount": 7
}
}
]
}
4.4 Scores → Codex extensions (no target equivalent)
DCAT-US and schema.org have no slots for predictive probability fields. Codex publishes a family of pre-computed civic scores (approval likelihood, sentiment intensity, regulatory risk) under the axiom: namespace as xsd:float [0,1] values, alongside structured tags for blocker conditions and contingent-approval graphs. Field-by-field semantics, computation methodology, and calibration metrics are documented in the commercial schema spec.
The axiom: namespace resolves at https://axiomcodex.io/ns/v1/ (planned, see AXC-3).
4.5 Spatial → OGC + H3 extension
| Codex field | OGC | Notes |
|---|---|---|
location (implicit in h3_index) |
OGC GeoJSON Point/Polygon |
Emitted in per-record JSON-LD as schema.org/GeoCoordinates. |
h3_index |
Codex extension | Exposed as axiom:h3Index (xsd:string, res-8 H3 cell). Not part of any W3C/OGC standard yet but widely adopted; documented in Uber H3 spec. |
5. FAIR compliance
| Principle | Codex status | Evidence |
|---|---|---|
| Findable | ✅ | Persistent record_id URNs, DCAT-US catalog.jsonld indexed for data.gov crawlers, listed on HuggingFace (AXC-5) |
| Accessible | ✅ | HTTPS fetch from R2-backed stream.codex.axiomlayer.io with signed URLs; MCP server for machine access |
| Interoperable | ✅ | JSON-LD with shared vocabularies (DCAT-US, PROV-DM, schema.org); Parquet standard columnar format |
| Reusable | ✅ | SPDX license identifier, clear attribution (dct:publisher), versioned schemas with deprecation policy |
DCAT-US v3.0 conformance level: 3 (all mandatory fields + all recommended fields populated; extensions documented). Level 4 requires publishing as part of a federated catalog; currently out of scope.
6. Codex extensions (no target equivalent)
Fields where no DCAT-US / PROV-DM / schema.org slot exists are published under the axiom: namespace. The two cross-cutting extensions are:
axiom:h3Index— H3 resolution-8 spatial keyaxiom:entityUrn— Codex entity-resolution key
The full set of dataset-specific extension fields (predictive scores, structured blocker tags, conditional-approval graphs, language-feature signals) is documented in the commercial schema spec. Each extension lives at https://axiomcodex.io/ns/v1/{term}.jsonld with a dereferenceable definition (planned per AXC-3).
7. Catalog sidecar example
Abbreviated catalog.jsonld emitted alongside the Parquet distribution:
{
"@context": [
"https://project-open-data.cio.gov/v1.1/schema/catalog.jsonld",
"https://www.w3.org/ns/prov.jsonld",
"https://schema.org/",
{"axiom": "https://axiomcodex.io/ns/v1/"}
],
"@type": "dcat:Catalog",
"conformsTo": "https://resources.data.gov/resources/dcat-us/",
"dataset": [{
"@type": "dcat:Dataset",
"identifier": "urn:axiom:codex:dataset:civic-intelligence:v1.0.0",
"title": "Axiom Civic Intelligence Dataset",
"description": "Normalized council proceedings, permits, zoning hearings, and civic filings across U.S. metros, with LLM-extracted entities, blockers, and predictive scores.",
"publisher": {"@type": "Organization", "name": "Axiomancer Labs LLC"},
"issued": "2026-04-16",
"modified": "2026-04-16",
"keyword": ["civic", "zoning", "council", "permits", "litigation-risk"],
"accessLevel": "public",
"license": "https://creativecommons.org/licenses/by/4.0/",
"spatial": "United States",
"temporal": "2020-01-01/2026-04-16",
"distribution": [
{
"@type": "dcat:Distribution",
"title": "Parquet snapshot",
"mediaType": "application/vnd.apache.parquet",
"accessURL": "https://codex.axiomlayer.io/datasets/civic-intelligence/v1.0.0/full.parquet"
},
{
"@type": "dcat:Distribution",
"title": "Markdown-KV LLM text view",
"mediaType": "text/markdown",
"accessURL": "https://codex.axiomlayer.io/datasets/civic-intelligence/v1.0.0/llm_text.tar.gz"
}
]
}]
}
8. Deviations from standards
- DCAT-US does not define a
h3_indexequivalent. Codex publishesdct:spatialas GeoJSON polygon (H3 cell boundary) andaxiom:h3Indexas the canonical string. - schema.org
EventstartDate is date OR datetime. Codex always uses ISO 8601 datetime with timezone. - PROV-DM requires a
prov:Entityper versioned artifact. Codex collapses the pipelineprov:Activityentries into the compactprovenance[]array on the record; full PROV-DM expansion is available via an auxiliary/provenance/{record_id}endpoint.
9. Changelog
- v1.0 (2026-04-16) — Initial crosswalk. Level 3 DCAT-US v3.0 conformance.
axiom:namespace placeholder — dereference endpoint pending AXC-3.