Loneliness is kind of a conundrum, when you dig into the data. You go into an exercise like this with some assumptions. And those assumptions are quickly put to the test. As just one example, ZIP code 94720 is, by most measures, one of the least isolated places on the continent. It is the University of California, Berkeley: 7,500 residents packed at more than 7,500 people per square mile, surrounded by libraries, dining halls, cafes, clubs for every conceivable interest, and tens of thousands of classmates within a ten-minute walk. When the CDC’s modelers estimate how many adults there feel lonely, the answer is 51 percent.
That’s over half! And it in an area where no one is really technically ever “alone.”
As we dive into the data you’ll see that population density has nothing to do with loneliness. It has more to do with age. In short, the loneliest addresses in America are full of nineteen-year-olds. The least lonely are full of seventy-five-year-olds. And almost everything else we assume about where loneliness lives gets at least one detail wrong.
Where is everybody lonely?
The measure here is specific, so one sentence on it: through the Behavioral Risk Factor Surveillance System, the CDC asks adults how often they feel lonely, and its PLACES project models the share answering “always,” “usually,” or “sometimes” for every ZIP-code area in participating states. Nationally that’s about 33 percent of adults — a third of the country lonely at least sometimes. The interesting part is the spread, from 19 percent in the calmest corner of the map to over 50 in the worst.
Among ZIP codes with at least 4,000 residents, here is the top of the list. See if you notice the theme:
It does not stop at ten. Auburn, Kent State, UC San Diego, UC Santa Cruz, Bowling Green, Binghamton, Liberty, Michigan State, Cal Poly, Virginia Tech, James Madison, Minnesota, UC Santa Barbara, Syracuse, Maryland, Ball State, Ole Miss, UConn, Wisconsin, Grambling State, Iowa State, Notre Dame, Delaware, Stanford, Clemson. Of the fifty loneliest sizable ZIP codes in the United States, roughly thirty-five are dedicated university campus ZIPs. The American college campus — the most socially engineered environment we have ever built, with orientation weeks and intramurals and resident advisors whose literal job is community — is, by the CDC’s numbers, the national capital of feeling alone.
Anyone who remembers their freshman year may find this less surprising than the architects of orientation week would. Loneliness, the researchers keep telling us, is not the absence of people. It is the absence of being known. A campus delivers ten thousand acquaintances and asks you to rebuild, from zero, at eighteen, every relationship that previously made you legible to anyone. The data suggests the gap between those two things is the widest it will ever be in an American life.
The other two clusters nobody talks about
Strip out the campuses and the top fifty has two more themes, both of them uncomfortable.
The first is military bases. Camp Pendleton (47.7%), Marine Corps Air Station Miramar (48.1%), Naval Base San Diego (46.8%), Naval Station Norfolk (46.5%), Twentynine Palms (46.4%) — every one of them ranks among the loneliest places in the country. The mechanism rhymes with the campus story: young adults, far from home, dropped into an institution full of peers and stripped of their existing ties, on rotations designed to move them again the moment roots form.
The second cluster is South Los Angeles. ZIP codes 90001, 90002, 90003, 90007, 90011, 90037, and 90059 — Florence, Watts, Vermont Square, South Park — all sit between 46 and 47 percent. There is no university here and no naval station. These are some of the poorest urban neighborhoods in California, and their loneliness has a different texture in the data: the campus ZIPs also post some of the nation’s highest depression estimates (Bowling Green’s campus hits 41 percent), while the South LA ZIPs report depression below the national average even as loneliness runs fifteen points above it. Whether that gap reflects genuine resilience, cultural differences in what people will tell a phone surveyor, or a healthcare system that diagnoses depression mostly in people who can afford to have it diagnosed — the data cannot say. What it can say is that two very different Americas report the same ache.
The least lonely place in America is a farm town in Oregon
Now run the list bottom-up. The least lonely sizable ZIP code in the United States is 97913 — Nyssa, Oregon, an onion-and-potato town on the Idaho border, where 19 percent of adults report feeling lonely. After it: Hilton Head Island (22.3%), the Sun City retirement community outside Bluffton, SC (22.9%), Isle of Palms, SC (23.6%), Pawleys Island, Sedona, Rehoboth Beach, Long Beach Island, Sun City West, Arizona — plus a belt of settled, prosperous suburbs like Leawood, Kansas and Mequon, Wisconsin.
Look at who lives in these places. Retirees, mostly. Long-tenured homeowners. The exact demographic every loneliness headline warns us about. Yet the towns where America’s old people actually live post loneliness rates barely half of what the campuses post. The popular story — youth as the social golden age, old age as the lonely epilogue — is upside down in the data. People in their golden-years communities have had fifty years to accumulate the thing campuses can’t issue at orientation: relationships with tenure. They also chose those communities, often specifically to be near people like themselves, which is its own quiet advantage.
(A modeling caveat belongs here, and it cuts both ways: these are estimates built partly from demographics, so places full of young adults will skew lonely because young adults nationally report more loneliness. That is not a flaw in the finding. That is the finding — being young in America right now carries roughly double the loneliness load of being old, and the map simply shows where the young people are.)
We tested the urbanist’s favorite fix. It failed.
Our Neighborhood Wellness Score project maps every gathering spot in 18,357 ZIP codes — parks, gyms, grocery stores, markets, cafes, restaurants. The “third place” theory of loneliness says these are the cure: build places to gather and people will gather, and gathering becomes belonging. So we crossed the two datasets, 13,443 ZIP codes where both exist, expecting to find loneliness thinning out where third places cluster.
The correlation runs the other way. ZIP codes with more gathering places are lonelier — and not just because cities have both more cafes and more lonely people. Even comparing only dense urban ZIPs against other dense urban ZIPs, the fifth with the most third places averages 2.6 percentage points more loneliness than the fifth with the fewest. The 31 ZIP codes in our data with zero gathering places of any kind — no park, no gym, no grocery store — are barely lonelier than the national average (34.5 vs 33.4 percent).
Causality could run several directions at once: lonely-prone people — the young, the single, the recently arrived — move to dense, amenity-rich neighborhoods, partly in hope of fixing it. But at minimum, the data refuses the comfortable version of the story. ZIP 94704 — downtown Berkeley, Telegraph Avenue — holds 385 places to gather, more than almost any ZIP we track, and sits in the loneliest percentile of the country at 47.1%. You can be surrounded by third places and have no place that knows your name. Infrastructure, it turns out, is the easy half of community. Berkeley has every amenity on the planner’s checklist. Nyssa, Oregon has onion fields and people who have known each other for forty years.
What this data can and cannot say
These are modeled small-area estimates, not door-to-door surveys — the CDC builds them from BRFSS responses and census demographics, and individual ZIPs can miss. The threshold is generous (“sometimes” lonely counts), which is why the national number is a third rather than a tenth; estimates of severe loneliness run lower. Not every state fields the survey module, so 8,000-odd ZIPs have no estimate. And a foot-traffic dataset can count gathering places but cannot see gathering — it knows where the cafe is, not whether anyone at it is known.
What survives every one of those caveats: the three loneliest kinds of places in America are the dorm, the base, and the disinvested urban core — institutions and neighborhoods full of people in transit, transplanted, or trapped, at densities that make the loneliness invisible from the outside. If the loneliness epidemic the Surgeon General declared in 2023 has a geography, it is not the empty front porch. It is a crowded room.
About this data
Loneliness, depression, and related estimates come from the CDC PLACES project (2025 release), which models Behavioral Risk Factor Surveillance System survey responses to ZIP-code areas; the loneliness measure counts adults who report always, usually, or sometimes feeling lonely. We analyzed the 24,234 ZIP-code areas with published estimates; ranked lists are restricted to ZIPs with 4,000+ residents. Gathering-place counts come from our proprietary Neighborhood Wellness Score dataset covering 18,357 residential ZIP codes (mapped venues and verified visit data, November 1, 2025 – April 30, 2026 window). Density comparisons use CDC/ATSDR Social Vulnerability Index 2022 population and land-area figures. Check any ZIP’s full profile with the Neighborhood Wellness Score.
Cory Jones has spent 20 years in media and publishing and now leads the data and editorial operation behind The Greenest. He built and maintains the site’s Neighborhood Wellness Score — a proprietary dataset scoring food access, walkability, air quality, healthcare access, and health outcomes across 18,356 U.S. ZIP codes, drawing on CDC PLACES data, EPA air quality readings, and verified location intelligence. He is a huge fan of Rancho Gordo beans and tries to work out more than he actually works out.
Is your neighborhood healthy?
Get a free report on your tap water, air quality, walkability and more
