2>4
The data engine

Federal data, not estimates dressed up as facts.

Tech Pathways doesn't scrape job boards or buy pre-packaged "career data." Every number you see — wages, target occupations, apprenticeship pathways — comes from a federal API or research-grade public dataset, joined together with explicit provenance and confidence tiers.

Programs covered
14,360
Wage-mapped
97.7%
Job-mapped
97.7%
Apprenticeship-mapped
97.3%

By the numbers

Built 2026-05-11
Schools2,15457 states + DC + PR
Programs14,360engineering tech, mechanic/repair, precision production, transportation
Occupations20041 flagged Bright Outlook (high-growth)
Wages197BLS OEWS May 2024 — national + state
Apprenticeships125RAPIDS occupation codes mapped into scope
Employers (seed)1735 states geocoded

How a single program becomes a career.

Every program record carries a 4-digit CIP code from Scorecard. From there, the engine joins outward through three federal crosswalks until you have wages, target jobs, and apprenticeship paths on the same row.

  1. 1
    Program → CIPvia College Scorecard

    Every U.S. CTE program reports a Classification of Instructional Programs code. Scorecard publishes them at the 4-digit level.

  2. 2
    CIP → 6-digit CIP → SOCvia O*NET education crosswalk

    O*NET's education crosswalk fans the 4-digit code out to 6-digit specializations, then maps each to one or more Standard Occupational Classification codes.

  3. 3
    SOC → wagesvia BLS OEWS

    The Bureau of Labor Statistics publishes employment + percentile wages for every SOC × geography. We pull national + state and surface the local one when you're looking at a school in that state. (Currently MN; other states one config flag away.)

  4. 4
    SOC → apprenticeshipsvia RAPIDS via O*NET

    RAPIDS is the federal registry of registered apprenticeship occupations. O*NET's RAPIDS crosswalk links every SOC to its apprenticeship code(s) — the alternative path to the same wage.

What we won't claim.

No fabricated employer-program edges.

“Companies that hire from program X” is mostly unobtainable as structured data without a paid feed (Lightcast, $30-100K/yr) or a state PIRL data-sharing agreement. We don't make it up. Where we surface employers, it's because they're documented industry partners, registered apprenticeship sponsors, or federally-tracked entities.

Every record carries a confidence tier.

Tier A = federal API or research dataset. Tier B = scraped under explicit structured headers. Tier C = LLM-extracted from narrative. Tier D = inferred via crosswalk only. Tiers never silently merge in a query.

Wages are an occupation median, not a graduation outcome.

Median wage shown on a program is the BLS-published median for the target occupation — what a working professional in that field earns today. It's not a placement guarantee. Where Scorecard publishes graduate earnings (~21% of programs), that's a separate signal.

Geocoding is honest about its limits.

OSM Nominatim is great for small towns and city centers, sparse for industrial corporate parks. Where we couldn't pin a specific HQ, the marker falls back to city-centroid and the underlying record is flagged.

Ready to look at it?

Open the map, pick a school, click a program. You'll see target jobs, median wage, and apprenticeship pathways surfaced together — that's the engine in three clicks.