LEHD Commuter Flows — Ontology Crosswalk

Schema: pending (AXC-11 — schemas/lehd-flows/v1.0.md) Target standards: U.S. Census TIGER/Line + FIPS codes, OGC GeoJSON (RFC 7946), DCAT-US v3.0, DDI-Lifecycle (statistical metadata) Authority: U.S. Census Bureau, OGC, IETF Status: FAIR-aligned, Census-native field mapping Dataset SPDX license: Public Domain (Census source) → republished as CC0-1.0 (research) · commercial tiers add Axiom enrichment under proprietary license

1. Why Census TIGER + FIPS

LEHD Origin-Destination Employment Statistics (LODES) are U.S. Census Bureau products. Every consuming pipeline in academia, urban planning, and transportation expects Census-native identifiers — FIPS codes for states/counties/tracts/block groups/blocks and TIGER/Line geometries as the spatial reference. Codex preserves Census fidelity and adds H3 as an interoperability layer without breaking the primary identifiers.

2. Dataset-level (DCAT-US v3.0)

DCAT-US field	Value
`dct:identifier`	`urn:axiom:codex:dataset:lehd-flows:v1.0.0`
`dct:title`	"Axiom LEHD Origin-Destination Commuter Flows Dataset"
`dct:description`	"LEHD LODES v8 origin-destination commuter flows normalized to H3 resolution 8 cells with pre-computed accessibility indices."
`dcat:keyword`	`["lehd", "lodes", "commuter-flows", "origin-destination", "census", "labor-mobility"]`
`dcat:theme`	`["http://publications.europa.eu/resource/authority/data-theme/ECON", "...SOCI"]`
`dct:source`	`https://lehd.ces.census.gov/data/lodes/`
`dct:license`	`https://creativecommons.org/publicdomain/zero/1.0/` (research)

3. Field-level crosswalk → Census LEHD LODES v8

Codex field	LODES equivalent	Notes
`origin_h3`	derived from `w_geocode` (15-char block FIPS)	H3 res 8 centroid of block geometry
`destination_h3`	derived from `h_geocode`	same
`origin_block_fips`	`w_geocode`	Retained as Census-native identifier
`destination_block_fips`	`h_geocode`	Retained
`worker_count`	`S000`	Total jobs
`worker_count_lt30k`	`SE01`	Jobs with earnings < $1250/month
`worker_count_30to60k`	`SE02`	Jobs with earnings $1251–$3333/month
`worker_count_gt60k`	`SE03`	Jobs with earnings > $3333/month
`income_band`	derived from SE01/02/03	Codex-published enum: `low`, `mid`, `high`
`job_sector_naics`	`SI01`–`SI03`	Sector-aggregated counts
`metro_slug`	derived	Metro FIPS → slug
`lehd_year`	file name / metadata	e.g. `2023`
`normalization_version`	fixed	`1.0.0`

3.1 Codex extensions

Codex field	Notes
`accessibility_index`	Codex-computed accessibility score per origin-destination H3 pair. Methodology details available in the commercial schema spec.
`distance_km`	Centroid-to-centroid km between H3 cells
`h3_neighbor_rank`	k-ring distance from origin to destination (0 = same cell)

4. Spatial crosswalk → OGC GeoJSON

Per RFC 7946, each flow is a LineString Feature when queried spatially:

{
  "type": "Feature",
  "geometry": {
    "type": "LineString",
    "coordinates": [
      [-87.65, 41.88],  // origin H3 centroid
      [-87.62, 41.85]   // destination H3 centroid
    ]
  },
  "properties": {
    "origin_h3": "882664c2a9fffff",
    "destination_h3": "882664c2b5fffff",
    "worker_count": 218,
    "income_band": "mid",
    "accessibility_index": 0.042
  }
}

5. Statistical metadata → DDI-Lifecycle (brief)

For academic users, a DDI-L 3.3 StudyUnit sidecar describes:

StudyUnit.universe: "U.S. private-sector covered employment"
StudyUnit.coverage.spatial: "United States, 2002–current year"
StudyUnit.coverage.temporal: LODES year
StudyUnit.methodology: reference to LEHD QWI methodology paper
StudyUnit.variables[]: full S000, SE0n, SI0n list with Census-native definitions

Full DDI sidecar emitted at /datasets/lehd-flows/v1.0.0/ddi.xml.

6. FAIR compliance

Principle	Status
F	✅ (Census-native FIPS preserved + DCAT-US + DDI)
A	✅ (HTTPS Parquet + NDJSON)
I	✅ (LODES columns preserved; H3 layered on top)
R	✅ (CC0 research tier, documented Huff methodology)

7. Deviations

Noise infusion is inherited. LODES applies noise infusion to protect confidentiality; Codex does not attempt to reverse. All counts carry the LODES noise properties.
Block-level granularity collapsed. We publish H3 cell pairs, not FIPS block pairs, as the primary key. Block pairs are retained as supplementary columns for academic users but deduplicated at the H3 level by default.

8. Changelog

v1.0 (2026-04-16) — Initial crosswalk. AXC-11 schema spec pending.