The economist who ships
On the overlap between measurement rigor and shipping software — and why the two disciplines need each other.
People are sometimes surprised that a PhD economist hand-builds Spark pipelines. I think the surprise is the interesting part.
Economics taught me to ask a specific, unfashionable question of any dataset: what does this actually let you claim? Most measurement is worse than it looks, and most "insights" are artifacts. Engineering taught me the complementary discipline: a claim you can't put into production, observe, and audit is a claim you don't really hold.
The overlap is rarer than it should be. Plenty of engineers can move data at scale but can't tell you whether the number at the end means anything. Plenty of economists can reason about identification but can't ship the system that produces the number in the first place.
I sit in the overlap on purpose. It's why the systems I build — a platform that quantifies the entire federal regulatory corpus, a live pipeline correlating public discourse with economic indicators — are about trustworthy data, not just big data.
(Seed post — replace with your own.)