Statecraft Ten Thoughts on Government Data
Mar 5, 2026
Violet Buxton Walsh, an immigration fellow who analyzes government immigration data, reads her essay and recounts building the OPT Observatory from SEVIS. She explains why government datasets feel impenetrable. She highlights major gaps, surprising errors, how forms reveal hidden fields, and why practitioner knowledge and creative restructuring matter.
AI Snips
Chapters
Transcript
Episode notes
Administrative Data Has Major Gaps
- Administrative government data often misses critical fields and events, not just by omission but because systems never collected them reliably.
- SEVIS lacked consistent departure dates and employer addresses, so DHS and researchers can't plainly know which visa holders remain or where students work.
When Data Looks Wrong It Often Is
- Apparent anomalies in government datasets frequently signal real underlying errors or collection changes.
- Example: the U.S. undercounted international students by ~200,000 in 2024 until a diligent user flagged the issue months later.
Read The Forms To Find Hidden Fields
- Use the actual government forms to discover what data a system contains and where it lives.
- Violet used knowledge of USCIS forms (I-129) via an immigration lawyer to find wage and status fields that enriched analysis.