Prepared by the Burning Glass Institute · Data through 2025 · Version 2026-09-25
The tool covers 187 detailed occupations that call for hands-on technical skill learned through apprenticeship, vocational training or extended on-the-job learning. Several federal surveys report some of these occupations together, so the tool shows them as 124 occupations in five groups: construction and extraction; installation, maintenance and repair; production; transportation and material moving; and engineering technicians. Every state and the District of Columbia are covered.
| Source | What it provides |
|---|---|
| BLS Occupational Employment and Wage Statistics | Employment and wages by state and occupation, 2019–2025 |
| BLS Employment Projections | National openings, projected growth, turnover and exit rates, 2025–2035; typical entry education, 2024–2034 |
| State long-term projections (Projections Central) | Openings and projected growth by state, 2024–2034 |
| Census American Community Survey (2020–2024) | Age, sex, part-time work, interstate migration, wages by education |
| Census Current Population Survey | Certification and license holding |
| US Department of Labor apprenticeship records (RAPIDS) | Active apprentices, starts, completions |
| Department of Education: IPEDS, College Scorecard | Credentials awarded, completion rates, earnings after completion |
| Perkins V state reports | High-school career and technical education (CTE) |
| BEA GDP by industry | Value added |
| Burning Glass Institute career-history data | Where graduates and apprentices go on to work, and how long workers stay in the trades |
| Burning Glass Institute Job Postings Database | Skills employers ask for |
Sources are used at their most recent reliable release, so vintages differ. Figures are shown as published and rounded for display. Where a figure rests on a small sample it is faded or labeled low precision; where it is not published, or too thin to stand on its own, it is left blank rather than shown as zero.
Employment counts wage and salary jobs located in the state, and wages are median annual pay without adjustment for inflation. National occupation figures come from BLS Employment Projections for 2025 and also count the self-employed, so they usually exceed the sum of the states, though not for every occupation. Where an occupation combines several detailed jobs, its wages are employment-weighted averages of those jobs.
Where BLS did not publish a year, the trend line fills it from neighboring published years; filled points are marked. Growth is the average yearly change between the published 2019 and 2025 figures, so it does not use filled points, is blank where either year is unpublished, and is labeled low precision for small occupations. The national trades median wage is the employment-weighted median of the state and occupation medians.
Age, sex, part-time work and industry come from the American Community Survey, pooled over 2020–2024. Part-time means usually working under 35 hours a week. Survey figures are graded by how many people they rest on, and a profile built on very few respondents says so.
The retirement rate is an estimate rather than a measurement: the share of workers aged 55 and over, spread evenly over ten years. Estimated retirements a year use the same rule.
Estimated retirements a year use the rule above. Counted entrants a year are credentials of a year or more plus completed apprenticeships, recorded nationally for that trade. They are blank for trades outside the set where both routes can be counted end to end.
Of people who started in the occupation in that state and are observed working five years later, the share working in any skilled trade, from online career histories. Where samples are small, state figures lean toward the national figure for the occupation, so most of the variation is between occupations rather than between states.
Openings and projected growth are the states' own official projections, not a Burning Glass forecast. Openings a year combine new jobs from growth with jobs left by workers leaving the labor force or moving to another occupation. Where a state does not project an occupation, openings are estimated from national rates and marked; where a state projects an occupation only inside a combined line, that line is split by employment and marked.
Net new workers needed removes the openings filled by people switching in from other occupations. A training system does not have to supply those workers, so this is the figure the tool compares training output against. Modeled openings apply national turnover rates to 2025 employment and leave growth out. The openings, net-new-workers and modeled-openings figures on this panel all follow the 2024–2034 state projection round; the national occupation figures follow the 2025–2035 round, so the two sets are not on the same basis.
The index scores the structural pressure on filling an occupation from nine indicators, including workforce age, new entrants, part-time work, licensing, AI exposure and projected growth. It ranks each occupation among all occupations in its own state, trades and non-trades alike, and a state scores between 147 and 522 occupations, so compare percentiles across states, not ranks or scores. It measures risk rather than current vacancies. Occupations with fewer than 20 survey respondents in a state are not scored. An occupation ranked 1 faces the most pressure in its state. The index is an occupation reading; states are not ranked on it, and the state-level comparison is the supply and demand indexes below.
Each occupation is placed in one of four typical entry routes: registered apprenticeship, a college or CTE credential, open entry, or on the job with no counted training route. "Mixed" means the detailed occupations within it enter differently. Separately, the tool records whether a college program, a registered apprenticeship and a high-school CTE program lead to that occupation in that state. "Unknown" means reporting was not complete enough to say a route is missing.
Credential supply counts all college credentials awarded in trade programs in 2024, including certificates of under a year, assigned to occupations by where each program's graduates are observed working. Dividing it by the net new workers needed gives credential coverage, which is low in most trades by design, because most trades are entered through apprenticeship or on the job rather than through college. A low ratio is not a measure of unfilled demand. Tracked coverage counts credentials of a year or more plus apprenticeships completed in 2024, and runs only on the occupations where both routes can be counted, which carry about a quarter of the need.
College completion is the share of students in trade programs who complete within four years, estimated from each college's cohort results rather than from individual students. There are two versions. The national figure, 48.3%, covers all providers, including certificate-only schools, and weights each college by its trade awards. The state map and the occupation figures cover degree-granting colleges only, whose national rate is 39.9%. Do not average the state map into a national rate: its median state, 40.9%, is neither figure. An occupation's rate is the completion rate of the trade programs whose graduates go on to work in it, not of a program named for it.
Where people are lost covers not finishing college, canceling an apprenticeship, and not reaching a trade after graduating. The last of these covers graduates at every degree level, including bachelor's and graduate degrees, so it is not comparable with the trade capture rate. Apprenticeship cancellations are shown only for states whose apprenticeship completion rate is shown.
The trade capture rate is, among graduates of sub-baccalaureate trade programs who are observed working five years after graduating, the share working in a trade. It is measured on earlier graduating classes and applied to programs' 2022–2024 graduates. Career histories under-count trade work, so the rate is more likely understated than overstated.
Earnings are typical earnings a year after completing for graduates working and not enrolled, from the latest College Scorecard release. Where a state's own program figures are too sparse the national figure is used, and figures below the reliability standard are left blank. High-school CTE counts students concentrating in trade-related clusters in 2023–24, and the trade share is shown only for the 20 states whose reporting supports it.
Apprentices per 1,000 workers counts active registered apprentices against employment in the occupations usually entered through apprenticeship, adjusted for states that do not record every apprentice's occupation. The completion rate is the share who completed rather than canceled among apprentices scheduled to finish in 2021–2024, from Department of Labor figures. Apprentices still training a year after their scheduled end are left out, so the rate sits slightly above the one the Department of Labor publishes. Six states and the District of Columbia are blank: New Mexico publishes no figure, and in the other six the published figure covers under a quarter of the state's apprentices. Completion rates depend partly on which trades a state trains, so compare states with similar trade mixes. The maturity score combines a state's position on apprentice numbers with its position on completion, so a state has to do well on both. For the seven without a completion rate, the score uses the national average completion rate, so it reflects apprentice numbers only. Completions a year count registered apprenticeship completions in the trades in 2024; estimated completions at current volumes multiply current starts by the national completion rate.
The share of registered apprentices who started between 2016 and 2022 and were working in a trade within five years of leaving the apprenticeship, from career histories. The range shows statistical uncertainty, and the completer range estimates the rate for those who completed their program. Where the completer range rests on too little federal completion data it is faded; read the overall rate instead. Career histories under-represent union programs, so these figures are best read nationally and by trade group rather than compared between states.
Moves between states are yearly averages over 2020–2024, with occupation as reported after the move. The 2022 survey suppresses the prior state of residence of movers into and out of Connecticut, so flows into and out of Connecticut average the four years in which origins are identified, and suppressed origins are not counted as arrivals from abroad. The net migration rate is arrivals minus departures per 100 trades workers, shown beside the same rate for all workers so that a state's general pull can be separated from its pull on the trades. Each state's six largest corridors are drawn; corridors based on fewer than 20 surveyed movers are not.
The tool measures pressure on each state's trades with two indexes and the position a state holds on both at once. Both are state-level: there is no score for a single trade within a state.
Demand pressure has four parts, each a single measure: projected net new workers needed a year per worker, actual employment growth from 2019 to 2025, whether the state's mix leans toward trades that need many new workers nationally, and real wage growth on a fixed mix of trades. Wage growth carries half weight because it is the least precisely measured.
Supply position has four parts and seven measures:
The two indexes are the tool's six indicators measured at state level. Everything in demand pressure is growth: where the work is growing, the new-worker need the state's mix of trades implies at national rates, and the wage signal that goes with rising demand. Supply position spans entry (counted entrants), conversion (training throughput), retirement and replacement (age of workforce) and talent movement (migration pull). Perception is measured nationally and has no state measure.
Each measure is converted to the state's percentile among the 51, running 0 to 100 with the highest state at 100. A part's score is the average of its measures, and an index is the weighted average of its four parts. A measure a state does not have is left out of the average rather than counted as zero: six states and the District of Columbia have no apprenticeship completion measure, and the District of Columbia has too few graduates for a measure of graduates working in a trade. The District's training throughput is therefore its college completion rate alone; that one rate, from a small number of institutions, puts it first on the part, so read its standing there with care.
A score is a position among the 51, not a quantity: 80 means high in the ranking, not 80% of anything, and a state's score moves when other states move.
Relative position reads the two rankings together. A state at least five rank places above the middle on demand pressure and at least five below it on supply position shows as *high demand, low supply*; a state at least five places on the favorable side of either index shows as *neither*; the rest are *borderline*, because a state a place or two from the middle would change class on noise. The band is in rank places so the class can be checked from the two ranks. The gap, demand pressure less supply position, is also shown; it is rounded from the unrounded scores, so it can differ by one from the difference of the two displayed scores. The two indexes rest on different measures and different occupations, so neither the class nor the gap establishes whether a state's supply meets its demand, and neither is a count of workers.
Job postings in the Burning Glass Institute Job Postings Database are sourced from Revelio Labs and classified by BGI models. Skill lists come from those postings for the 54 occupations with enough postings to rank skills, and each list shows its top five. Not every occupation has all three lists.
A skill is listed only if, where it is ranked on growth, that growth holds up on a stable set of job boards; where it is ranked on pay, its premium is statistically reliable; the trade asks for it at least twice as often as the labor market overall; and demand for it is spread across the equivalent of at least five independent employers. Compare skills only within one occupation. The premium describes what employers advertise, not what learning the skill would add to a worker's pay, and postings omit the basics every employer assumes, so an unlisted skill may still be essential.
Value added applies the trades' share of employment in each industry to that industry's value added in 2025, giving the trades' contribution to GDP. Industries the occupational survey does not cover, such as farms and government, are left out, so the total is understated. The national total is divided between states and occupations in proportion to each one's wage bill, its employment multiplied by its median wage.