About
A public mirror of US H-1B salary data,
built for speed and honesty.
The US Department of Labor publishes every certified Labor Condition Application (LCA) it receives — the form an employer must file before sponsoring an H-1B, H-1B1, or E-3 worker. The raw releases are quarterly spreadsheets totalling tens of gigabytes, with no consistent employer naming, no canonical job-title taxonomy, and a schema that changes every few years.
h1b.report takes those releases and turns them into something a human can actually use: search any company in milliseconds, see every role and worksite, compare wages against the federal prevailing-wage tiers, and read trend lines back to fiscal year 2002.
How the data is normalized
- 1. Ingestion. A streaming pipeline ingests every quarterly OFLC release back to FY2002 — roughly 12 million certified records — without ever loading a full file into memory.
- 2. Validation. Each record is checked against a strict schema; rows that fail validation are quarantined for review rather than silently dropped.
- 3. Employer resolution. A three-layer pipeline collapses millions of employer name variants — exact federal tax-ID match, then fuzzy string similarity, then semantic embeddings — into canonical organizations.
- 4. Occupation classification. Free-text job titles are mapped to the Bureau of Labor Statistics SOC taxonomy, falling back to a fine-tuned transformer when no direct match exists. Low-confidence predictions are flagged for human review, not assigned silently.
- 5. Publication. Results are served as a fast, indexable web app — and as a documented JSON API for researchers, journalists, and policy analysts.
Who this is for
- Workers deciding whether an offer is competitive against actual filed wages for their role and location.
- Researchers studying immigration policy, wage-suppression hypotheses, or industry concentration of foreign labor.
- Journalists tracking which companies sponsor most, where wage trends diverge from the prevailing rate, and how the program shifts year to year.
- Employers benchmarking offered salaries against their peer set before filing.
What it isn't
h1b.report does not show approved visa petitions, individual worker names, or any data the Department of Labor itself does not publish. It is not affiliated with the US government, USCIS, or DOL. It is a re-presentation of public data under the same disclosures the government already makes.
Get in touch
Questions, corrections, or research collaboration: hello@h1b.report.