Where the applications are going
Every state, metro and ZIP3 region in our vocabulary is drawn below. A cell lights up once at least five separate contributors are counted in it. Everything else stays grey — and grey means “in the vocabulary, not in the response.” It never means zero.
Why so much grey?
Because we refuse to say anything about a place until five separate contributors are in it — and we enforce that when the number is written, not when it is shown. A cell below the threshold is never stored, so there is nothing to leak and nothing to reconstruct. Grey covers “nobody,” “one” and “four” with a single indistinguishable state, on purpose.
What a bubble is not
It is not a response rate, an interview rate, or a claim about your odds. At these tiers we publish volume only. A big bubble means a lot of applications went there — not that they went well.
How precise is a region?
A ZIP3 covers roughly 700 square miles. We never receive a street, a city, or a full ZIP — the truncation happens in the input field on your own machine. The metro tier is an overlay, not a national map: many regions belong to no metro. The state tier is the one with full coverage.
Why do three regions share a name?
Because a ZIP3 is a postal prefix, not a neighbourhood. Manhattan is 100, 101 and 102 — three prefixes, all named for the same post office city. 532 of our 918 region labels are shared with another code, so the code is shown first and the name second. We do not merge them into one “New York City” bubble: that would be a number nobody computed, and we only publish numbers we can defend. Zoom in — they separate.
What the numbers actually say.
The map shows where applications go. This shows how they do. Every rate below carries its sample size and a 95% confidence interval, every list is ordered by volume rather than by rate, and nothing measured on fewer than five contributors is published at all.