Methodology
How every number, score, and projection on CareerAtlas is calculated — and how much to trust it.
Every data-bearing figure on CareerAtlas is labeled with exactly one of four statuses:
Directly observed from an official or licensed source (e.g. a live BLS series).
Derived from official data with some interpolation or modeling.
A forward-looking projection, computed from a documented formula.
Deterministically generated placeholder data used to make the product usable before every connector is fully populated. Never treat as real-world observation.
Scoring formulas
version 1.0
Scores how accessible an occupation is to enter.
Score = weighted blend of entry-level opening share, inverse of typical required experience, inverse of degree-requirement strictness, inverse of skill-gap size, and number of geographic markets hiring for the role.
version 1.0
Optional combined score blending compensation, growth, accessibility, stability, flexibility, and transition potential.
Score = user-weighted average of Salary Opportunity Score, Momentum Score, Accessibility Score, Automation Safety, Remote Flexibility, and (inverted) Education Cost.
version 1.0
Estimates how much to trust a given data point.
Confidence = function of sample size, data recency, and whether the value is reported (highest), estimated, forecast, or simulated (lowest baseline confidence).
version 1.0
Estimates the financial return on an education path from cost, time, and post-graduation earnings.
Net Cost = Total Tuition + Fees − Earnings During School (if part-time work assumed). Break-even Year = first year cumulative (post-grad earnings − no-degree baseline earnings) exceeds Net Cost + forgone earnings while in school. 10/20-year return = cumulative earnings differential over that horizon minus Net Cost, expressed as a percentage of Net Cost.
version 1.0
Composite 0-100 Job Market Momentum Score for an industry, built from nine weighted, user-adjustable sub-scores.
Score = Σ(weight_i × subscore_i) over: employment growth, posting growth, salary growth, hiring velocity, layoff risk (inverted), skill demand growth, automation safety (inverted exposure), entry-level availability, geographic diversity. Default weights sum to 1.0 and are shown in the UI, which lets users override them.
version 1.0
Scores an occupation's compensation attractiveness.
Score = weighted blend of median total compensation percentile rank, 5-year salary growth, salary ceiling (p90/median spread), and cost-of-living-adjusted compensation percentile rank.
version 1.0
Projects future salary from current salary using transparent multiplicative factors rather than a black-box ML model.
Projected Salary = Current Salary × (1 + general wage growth)^years × occupation growth factor × industry momentum factor × experience factor × education factor × skill demand factor × promotion/transition factor. Conservative/Expected/Aggressive scenarios shift each factor down/center/up.
version 1.0
Scores how compatible and attractive a transition between two occupations is.
Compatibility = weighted overlap of skills, education requirements, experience level, and industry/job-family similarity. Opportunity, demand, and confidence are scored separately and shown alongside compatibility rather than blended into one hidden number.
Salary projections are estimates based on transparent, documented assumptions — not promises.
Education outcomes describe historical correlations, not causal guarantees. A degree does not cause a specific salary.
Transition scores describe historical compatibility between two roles, not a guaranteed career path.
Where data is insufficient, the product says so explicitly rather than guessing.