About us

How the numbers are made

In a minute: Every count on this site is our estimate from the Census Bureau's American Community Survey from 2020 through 2024. You select what you're looking for; we count how many matching people live in each of 387 US metro areas, and rank the 193 with at least 250,000 people and enough survey sample to answer. We use additional information from nationally-recognized datasets and weigh it all together to find your best city match.

Where does the data come from?

• People counts — the American Community Survey, about 3.5 million households a year, 2020–2024 combined, weighted the way the Census Bureau weighs them. Metro areas follow the government's own definitions: "San Francisco" means the whole metro area, not the city limits.
• Place stats — each from a nationally-recognized source, updated when that source publishes:

What does a "match" count?

Everyone in a metro area who fits your whole search — sex, age range, never married or previously married, and any education, income, race or ethnicity filters you set. That count is the dating pool, and it is the biggest part of the score.

Compatibility

Compatibility is how closely the people who match your search resemble the people who actually pair with someone like you, on age and education (and on background, if you include yours), where 100 is the US average.

The compatibility figure is built from real couples in the Census Bureau's household survey, using each partner's age and education, and their race or ethnicity only if you include yours. Across the whole country we measure how often each age gap and each education pairing actually occurs, compared with how often it would occur if single people paired at random; if you include your race or ethnicity, the racial and ethnic pairings are measured the same way. That pattern, adjusted a little for each city's own couples, is then applied to the single people who match your search in each city, so the figure says how closely they resemble the people who typically pair with someone of your sex, age and education, and background if you include it. It is an aggregate pattern from recent unions, not a prediction about any one person, and 100 is the US average for your search.

Your own sex, education and race or ethnicity stay in your browser and are never sent to us: the site works out the figure for every possible answer, and your browser shows the one that fits you. Your education is optional; leave it unset and the average for people of your age and sex is used instead. Your race or ethnicity is used only if you include it, and then only for this figure; without it, the figure uses no racial or ethnic pairing at all.

For a same-sex search the age gaps and the education pairings come from same-sex couples in the same survey; the racial and ethnic pairings are not used, even if you include your race or ethnicity.

What does balance compare?

Dating pool balance is a simpler number: all single men per 100 single women (or the mirror, if you're looking for women) in the age range you picked — never-married, divorced or widowed people, counted across the whole metro, before any education, income, race or ethnicity filter. It deliberately ignores your other filters so it means what it sounds like: the shape of the whole singles scene at that age, which doesn't move as you refine your search. It's shown per 100, rounded to a whole number, and only where both sides of the comparison have enough survey sample to be dependable. It is shown for information and is not part of the score.

What do the race and ethnicity boxes do?

They select who is counted as living in a city and matching your search — nothing more.

There are eight boxes — the Census Bureau's race and ethnicity categories, including "Two or more races" and "Another race" — and the boxes you tick are exactly who gets counted, with nothing added. Tick two groups and the count is the sum of those two groups. Unticking every box means the same as ticking all eight: no filter at all.

These boxes are about the people you're looking for. Your own race or ethnicity is a separate setting: it is used only if you include it, stays in your browser, and affects only the compatibility figure.

How does the score work?

You set what matters. The slider divides the people-side weight between pool size and compatibility; four controls — cost of living, social life, student life, weather — set how much each place stat counts. "Not much" makes a thing count a little, not zero, so nothing you deprioritise can silently vanish. Each city's stats are compared across the cities that can answer your search, weighted your way, and summed to a score out of 100. There are no hidden weights: the controls on the home page are the whole model.

Why do some searches leave cities out?

Because the data is a survey. A very specific search can turn up only a handful of matching responses in a smaller city, and below about a hundred matching respondents a count stops being dependable — so we leave that city out of your results and say how many were left out. Too few matches in the survey is not the same as too few people in the country: it usually means the search is narrow, not that nobody fits.

Every measure, every city

The full list of what this site measures — each statistic in plain words, with a preference-free list of every ranked city — lives on one page: What we measure.