If you work in workforce policy or economic development you already have access to county unemployment. BLS publishes Local Area Unemployment Statistics monthly, free, for every county in the country. Nobody needs to sell you that.
The difficulty is what to do with it.
The rate moves too slowly to be a warning
County unemployment is a lagging indicator. By the time a county's rate moves, the layoffs happened, the separations processed, and the claims filed. If you are trying to see where trouble is heading, the rate tells you where trouble already went.
It is also seasonally noisy in ways that swamp the signal. A county with a large agricultural or tourism sector swings every year on a schedule that has nothing to do with economic health, and comparing it to a metro county on raw rate alone produces a ranking that is mostly a map of seasonality.
Comparing counties is harder than it looks
A 6% rate in a county of 8,000 people and a 6% rate in a county of 800,000 are not the same event. Neither are they comparable across states with different industry mixes and different claim processing.
This is why ranking counties by raw unemployment produces a list that experienced analysts ignore. The ranking is dominated by the small and the seasonal.
What we actually built, and what we removed
Chandler holds 3,221 counties across 52 state and territory codes, each carrying **real BLS county unemployment — currently the 2026-06** period — alongside WARN filings and court signals, delivered through MCP so an AI agent or an analyst's tooling can query it directly.
**And here is the part worth saying plainly: this dataset used to contain more fields than it does now.** Earlier versions carried county attributes — industry dependency, workers at risk, wage and tax impact — that were derived rather than measured. Those fields have been nulled out across all 3,221 rows rather than shipped. A number that looks precise and is not sourced is worse than no number, particularly for research use where somebody will cite it.
What remains is what we can point at a source for. That is a smaller product than the one the earlier version appeared to be, and it is the honest one.
The score is a percentile — read it that way
The displacement score is a relative ranking, not an absolute measure of risk. A county in the 95th percentile is not "95% likely" to experience anything. It is ranked above 95% of other counties on the inputs we hold.
That makes the tails meaningful and the middle largely noise. If you are using this to decide where to look first, the top and bottom of the distribution are where the information is.
Pricing starts at $299. *(See note below — the tiers need work before this is a fair thing to put in a blog post.)*