EB-2 NIW · Profession Guide
EB-2 NIW for Data scientists: AAO Data, Denial Patterns & Evidence
Why data-scientist EB-2 NIW petitions fail at AAO when commercial work is not translated into public-interest framing.
Based on 6,362 real USCIS AAO decisions · Last updated May 2026
Short answer
Across 315 Data Scientist AAO decisions in our corpus, 5.7% were approved on appeal, 84.8% were denied, and 9.5% were remanded. The single most common denial reason for data scientists is “Commercial work ≠ national interest.” AAO rates are lower than first-pass USCIS rates because these cases were already denied at least once.
AAO outcomes for data scientists (315 decisions)
Read this carefully: AAO numbers reflect petitions that were already denied at least once and appealed. First-pass USCIS approval rates are substantially higher. Use these figures to understand which arguments USCIS finds insufficient at the highest scrutiny level.
Why data scientists get denied at AAO
Most common AAO denial reason in this bucket:
Commercial work ≠ national interest
AAO consistently treats data-science work in ad-tech, e-commerce, fintech, or martech as private commercial benefit. The same petitioner with the same skill set, working on healthcare risk models, climate signal extraction, or election-integrity analytics, has a much cleaner prong-one story. The framing decision matters more than the technical work for this bucket.
What strong data scientist petitions tend to include
These are the evidence types that recur in approved Data Scientist cases. Not every approved petition has all of them, but petitions missing several typically struggle at AAO.
- 1Public-interest data work (healthcare, climate, education, civic) — pro bono engagements count if documented
- 2Refereed papers or workshop submissions with citation evidence
- 3Open-source statistical / ML libraries with download or fork counts
- 4Speaking at refereed venues (KDD, JSM, useR!, PyData with selection committees)
- 5Letters from independent researchers, not your direct manager
- 6Patents or standards work on data-quality / privacy / fairness methodology
How data scientist cases fit the Dhanasar three-prong test
The Dhanasar framework asks USCIS to evaluate three things together: substantive merit, your positioning to advance the work, and whether waiving the labor cert makes sense on balance. Here is how the prongs typically frame for data scientists.
Prong 1 — Substantive merit and national importance
Replace "drove revenue" with "advanced public-interest analytics in [healthcare/climate/civic]". Re-frame, do not just translate.
Prong 2 — Well-positioned to advance the proposed endeavor
Citations, downloads, and conference selections do most of the work — your résumé is the weakest evidence here.
Prong 3 — On balance, waiver is in the national interest
Argue that gaps in the public-interest data workforce make labor cert impractical — cite BLS occupational data.
What approved Data Scientist profiles look like
Public-interest framing + at least one refereed paper or widely-used open-source artifact + independent expert letters.
This is a composite based on patterns across 315 AAO decisions — not any single case. Your specific profile may clear with less, or struggle with more, depending on framing.
Run a personalized Data Scientist case analysis
Aggregate data tells you what AAO has rejected for data scientists. A $10 ai case review tells you which of those failure modes your profile is closest to — prong by prong, with the five most-similar AAO cases pulled directly from the same 6,362-decision corpus.
One-time payment, no subscription. Greenway AI is a data + document-generation platform, not a law firm; nothing here is legal advice.