AI Career Radar

Methodology & data sources

Every score is reproducible from public U.S. government data. Here is how it is calculated — and what it does not claim.

Where the data comes from

  • O*NET Work Context — how work is done; drives the exposure proxy.
  • BLS OEWS — median annual wages by occupation.
  • BLS Employment Projections — 10-year projected employment growth.

Role Exposure Score

An occupation-level proxy on an absolute 0–100 scale: a weighted composite of three O*NET 31.0 Work Context descriptors — Degree of Automation (0.4), low Task Autonomy (0.3; the published autonomy value is inverted), and Importance of Repeating Same Tasks (0.3). Bands: Low 0–34, Moderate 35–49, High 50–100.

Skill Durability

Each skill gets a transferability score: how much growing, low-exposure occupations that treat it as a core requirement value it. A role's durability is the transferable skill capital it builds. Bands: Fragile below 55, Moderate 55–74, Durable 75 and above.

Pivot Path Planner

Every other occupation is scored as a target on four components — skill gap, retraining effort, pay change, and growth outlook. You choose the weighting; the ranking recomputes instantly in your browser.

Employment growth

Growth is the BLS ten-year projected employment change. Bands compare it with the all-occupation average of about 3.5%.

Limitations

  • Coverage includes 875 selected occupations, not the full SOC taxonomy.
  • Exposure weights are an initial calibration and will be refined.
  • Wages and growth reflect a point-in-time snapshot: BLS OEWS May 2025.
  • The model does not include employer adoption speed, local conditions, or an individual's task mix.