Opportunity has always depended on education. As artificial intelligence transforms the economy, it is reshaping the knowledge, skills, and credentials that create economic opportunity.
Opportunity Data is an independent research platform built on public data. We develop statistical infrastructure, decision-support frameworks, and transparent analysis that help researchers, policymakers, institutions, employers, and learners understand how technological change is reshaping education, work, and opportunity.
Building open infrastructure from public data.
Explore the data infrastructure
The 2026-27 list names 571 occupations covering 15.5 million of the state's 18.2 million jobs. Measured against state wage data, 51 percent of California's employment is both listed and below the living wage, and the federal earnings test passes every matched occupation at full-time hours. The gates that count are program quality and labor-market alignment, and colleges hold both.
The labor market is changing faster than the systems built to measure it. Administrative records, education data, and emerging A.I. indicators already contain signals about how opportunity is being reshaped, but those signals sit in separate systems designed for compliance, reporting, and institutional use, not for public understanding.
Opportunity Data turns those public resources into indexes, crosswalks, state-by-state implementation records, and interactive explorers that anyone can use and reuse. Rather than producing one-time reports, we build open infrastructure from public data. Opportunity Data is not tied to a vendor, membership organization, or predetermined implementation model.
Four interactive components, each grounded in public data sources and reproducible end-to-end.
Which occupations are most exposed to A.I. automation, and which academic programs feed into them? Three-dimensional scoring across digital intensity, human interaction, and physical work. 696 occupations, 1,806 programs.
How A.I.-exposed is each state's mix of graduates? A completion-weighted map of all 50 states and D.C., built from 5.3 million IPEDS 2023 completions and the A.I. Exposure Index, with credential breakdowns and a findings brief.
From eligibility screening to state implementation. Approximate the federal earnings test program by program, track each governor's progress in standing up the approval pipeline, and compare the occupation lists states are approving.
What graduates actually earn at 1, 5, and 10 years, drawn from Census Bureau administrative wage records. Trajectories, within-program variation, and where suppression limits what we can see. 1,100+ institutions in 37 states and DC, three credential levels.
How 73 bachelor's fields compare on underemployment, unemployment, early-career pay, and the number of graduates, mapped together so you can see both the odds and the scale. New York Fed outcomes paired with IPEDS completion counts.
The datasets behind every component are free to download with a citation. Paid products add what the free layer does not hold: back editions of the State Implementation Record, state-level nuclear workforce reports, and a program portfolio review for one college. Custom research and state reviews are quoted through Work With Us.
Every article, dataset, and visualization contributes to a larger body of open statistical infrastructure for understanding education, work, and opportunity. Interactive analyses paired with published op-eds in national and trade press.
Unemployment among new graduates in the most AI-exposed tenth of majors nearly doubled between 2022 and 2024, computer science enrollment fell by an eighth this fall, and the federal earnings test cannot score a cohort that graduated after ChatGPT before 2029. Public survey, Census, and Clearinghouse data read through the Opportunity Data AI Exposure Index and the rule text.
View analysis →Every state's certificate portfolio measured against two facts the federal Workforce Pell test collapses into one: whether a field's completers clear the national median wage, and whether their earnings keep climbing. A map, all 51 portfolios, and the four buckets that divide 1.02 million certificates.
View analysis →Nine entry-level credentials sort into four tracks with almost no bridges between them. BLS OEWS wage bands for every rung and Census PSEO graduate trajectories show which track stacks (nursing) and which stop where they start.
View analysis →Where the teacher-preparation ladder steps up. PSEO earnings for graduates of CIPs 13.10 (Special Ed), 13.12 (Subject Area/Secondary), and 13.13 (Levels & Methods/Elementary) at the associate's and bachelor's levels, paired with NY Fed underemployment and BLS K-12 teacher wages.
View analysis →Below the bachelor's, the child development credential ladder is flat: a certificate and an associate's earn the same. PSEO earnings for Human Development and Family Studies graduates, paired with BLS OEWS wages and MIT Living Wage data across states.
View analysis →Colorado-only cut of the national ladder. PSEO earnings for graduates of 16 Colorado colleges in Human Development and Family Studies, paired with the Colorado Department of Early Childhood's published lead-teacher vs. kindergarten-teacher wage comparison and Child Care Aware of America's 2024 Colorado prices.
View analysis →Do bachelor's degrees from community colleges deliver comparable labor market value? Using Census Bureau earnings data at 1, 5, and 10 years post-graduation, CCB graduates match or exceed traditional BA earnings in most fields by Year 5.
View analysis →
"It requires taking the word 'projected' seriously."
"One federal statute, a nineteenfold difference in how wide the door opens."
"Colorado solved access first. It never solved the workforce behind it."
"The jump is not a skill jump. It is a payer jump."
"Short-term credentials are now the most-cited emerging model in every sector surveyed."
"Community colleges aren't just sitting in a safe harbor. They're absorbing the students trying to get out of the weather."
"The value-added earnings test is not the obstacle most people think it is."
How the data infrastructure works. Data sources, formulas, scoring, and the limits of what the data can show. Neutral documentation, not commentary.
How the AI Exposure Index is constructed: variable selection from O*NET, dimension scoring for digital intensity, human interaction shield, and physical anchor, and the crosswalk from occupations to academic programs.
Read methodology →Data sources, formula, population filters, cohort pooling rationale, and the five key approximations behind the Value-Added Earnings Test.
Read methodology →What the Post-Secondary Employment Outcomes data can and can't tell us about program value. UI coverage gaps, small-cell suppression, and non-random missingness, layered out.
Read methodology →