There’s been a good deal of recent attention to Seattle’s continued growth spurt. The Upshot column in the New York Times points out that we’re also one of the few cities that is growing denser as we add population. In fact, Seattle is already cited as the 8th most dense of the 50 most populous U.S. cities. I’ll expand on that last fact in this post – hopefully giving some context for what our current state of density means relative to the other large cities of the U.S.

Two questions arise naturally: What is a “large” city? And how should density be measured? Here, I’ll define a “large city” as one with at least 100,000 residents. Such cities are in the 99th percentile of population for all incorporated places in the U.S. – so that seems sensible. As for density, I find the population-weighted density metric to be more informative and interesting than the usual “population divided-by area” measure. Population-weighted density measures the density at which the average person resides and is less sensitive to the amount of vacant land within city boundaries. For an excellent example of why one might prefer weighted density, see Honolulu, Hawaii. The traditional density is about 6,000 ppl/sq. mi., but the weighted density is closer to 25,000. That difference is like suburban Renton vs. Lower Queen Anne, so it is significant!

How does Seattle stack up when it comes to weighted density? To find out, I pulled census block group level population estimates for all U.S. cities with over 100,000 residents from the 2015 American Community Survey. In all, I calculated weighted densities for about 300 cities. Here’s what the distribution of those densities looks like:

The most common density is around 5,000 ppl/sq. mi. and is exemplified by fast-growing, sprawl-y Colorado Springs, Colorado. At about 13,500 ppl/sq. mi., Seattle manages to make the 90th percentile of the distribution – which means it is denser than 90% of the cities in my sample. And New York City is by far the king at almost 75,000 ppl/sq. mi.

To give a better sense for what is going on in that distribution, here’s a map for cities with more than 250,000 residents (there are too many cities to display at the 100k threshold):

Not surprisingly, high density is largely found on the coasts (and Chicago). The more typical American cities, density-wise, are in the south and the midwest. The political implications of that statement are obviously a subject of intense discussion these days. But since I’m not Richard Florida and I don’t have graduate students to run demographic regressions for me, I’ll digress and just display the top 20 cities in that map by weighted density:

A few words on the “Uncertainty” column: city boundaries don’t always line up with census block group boundaries, so my weighted density calculations are within +/- the reported percentage of the correct number.

Seattle manages to crack the top 20 – and we should be happy about that. However, the relatively high densities (near 20,000 ppl/sq. mi.) of the major Los Angeles metro cities prove that increasing density alone will not save the Puget Sound region if our growth patterns are car dependent.

53 Replies to “Seattle is Denser Than 90% of Large U.S. Cities”

  1. It would be helpful to ELI5 the weighted density measurement. Most people can easily wrap their head around ppl/sqmi density measurement, but the weighed density is new and needs explanation.

    Also, thought I am not Richard Florida, if you’d like help, I can run the regressions for you. A priori, I’m envisioning region, pop-change-since1900 (aka old vs new cities), % of metro pop in city pop (gets at the difference on small bound cities (Seattle, SF, Detroit) vs cities that are a much bigger part of the metro (e.g. Houston), pop demographics (race, age, income). But, I am sure there is lit that helps up figure out more interesting predictors…

  2. Car dependency. I think to solve that we need to build more than just 300 – 400 ft2 apartments. Maybe something couples and families could fit into, and that have family friendly amenities and spaces?

    1. Really good point, Josh. Pretty much like clothes or shoes- design skill, workmanship, and creativity count for more than square footage of material. And can be designed by people using modern machines, not other way around.

      “Car dependency”? Like food, more than one definition. Is Dependency like the need for a healthy diet that tastes good? Or like addiction to fat and sugar- stats say the latter is worse than meth, for addictive power and horrible results.

      Eastern Pierce County looks like second one. Giveaway is that like with meth addiction, the chemical itself creates a physical and mental yearning for a result that only yields more misery. Requiring more methamphetamine.

      For remedy, I think we need to start designing and building the development equivalent of model homes. Attractive places to live with living spaces carefully designed to be compact without feeling cramped.

      Think well-tailored suit. Having tailors in the family, I don’t buy the idea that good tailoring costs too much. Talent, skill, and practice delivers a lot of style for the money. And my family’s best-known Artist often sits on juries for public art. His take on just about everything bad: “It’s lazy!”

      “Privacy” argument, any current subdivision proves worst of all worlds. Large empty spaces are perfect for undetected inter-neighbor spying. Even before you could buy a drone at a mall camera store. But arrangement also sees to it that nobody knows if that moving van is delivering more bad furniture, or removing your last good stuff.

      Good use of space creates both privacy and mutual protection. “Eyes on the street” is opposite of spying. More like conversation and coffee while Watching the Block for trouble. Movie suggestion:”Citizen Jane”, showed how by destroying healthy density, sprawl directly created worse threats to freedom than crowding and dilapidation.

      Ballard didn’t die of high rents per se. Killer was the loss of the small-industrial economy that gave so many average working people enough of an income to afford to live in a freer and more comfortable neighborhood than they’d now get for any money in less dense places.

      But. The new breweries, distilleries, and roasting facilities only became possible when modern electronics rendered their giant mass producing forebears obsolete. “3D Printing” is excellent example of new industrial truth:

      Notice the three year old on Mom’s lap working the i-pad before he or she can talk.
      From age four on up, people can now design and create quality products for home use or sale. While ceasing from babyhood to hate math.

      And instead of a thousand workers impersonating machines ’til they got killed or dismembered, same thousand workers can each be researching and designing same number of different things.

      The Eiffel Tower and the Space Needle have three things in common. Excellent design and construction. Artistic beauty. And intended to be temporary structures for Exhibitions. Which I really think are now this region’s greatest need for both land use and transit.

      Skytrain itself, designed for Vancouver “Expo” (name says it) in 1986, was probably itself pioneered by our own Monorail 24 years earlier.

      At this point of development every major technological change has been “kicked off” with an Exhibition. Main purpose is not for entertainment or sales, but to introduce life-sized working examples of new ways of life.

      Like with our train-boarding problems, main obstacle to necessary transit-orientation is that most local people have never seen it. So whatever the necessary legal, financial, and political arrangements, we need to build the exact equivalent of present large scale developments. A way of life where transit serves the purpose now filled by automobiles.

      Including not only the homes we’re talking about, but small industries where people can earn enough to afford them, and also schools to learn to run them. Same tech world that brought beer and coffee back can do the same with the rest of Ballard. At worst, we’ll have a dozen beautiful parks with a quarter century start on future transit design.

      Being former monkeys, spirit, curiosity, and creativity are genetic dependencies for us. Can’t lose them, so might as well use them. When we can’t, we get stressed, and start to screech, bite, and eat the ends of our own tails.

      Mark Dublin

  3. What strikes me is how quickly it drops off, so that even being in the top 90% is not that dense. A comparison with foreign cities would show how distorted and outlying US “normal” is. I’d like to see Canadian cities on this same scale. European cities may not be a fair comparison because of their much longer development history and their historical status as defense strongholds, and Asian cities may have too many cultural differences, but Canada has the same development timeline we do and a similar culture so the differences are purely policy choices. I suspect most Canadian cities would be in the top ten.

    Car dependency makes the biggest difference, because every parking lot or garage entrance or 5-lane road is dead space that people have to walk past and can’t live in or work in or recreate in. LA and San Jose are claimed to have high density but what they really have is a more uniform density citywide that’s capped by parking minimums. Seattle has more wide-ranging differences in density.

    And those wide differences in density are where viewing cities as a whole can feel misleading. Seattle is a very different place in the central core compared to most of the area east of 15th Ave NE, north of 85th Street, and south of Columbia City. They’re almost different cities, and people often prefer one or the other. If you excluded these outer areas Seattle’s density would be higher. and conversely Outer-Seattle would rank with Dallas or the midwest. Whereas LA and San Jose with more uniform density would remain where they are.

    I’ve followed Richard Florida loosely over the years and I was glad to see he was a cofounder of CityLab so I’d see his work more often. I followed the link above and read his most recent articles and found some interesting quotes.

    “If populism and Trump reflects the anti-urban backlash of the right, [NIMBY] attempts to limit high-tech development in urban centers reflects an anti-urbanism of the left.” (article)

    He’s also written a book “The New Urban Crisis” which I’m off to the bookstore to get. This article outlines his five solutions for decreasing urban inequality and spreading prosperity to more cities. Some of them will be familiar to STB readers, but he makes a good argument for them that’s worth reading.

    1. Yeah, I wish it was a lot easier to compare American cities with Canadian cities. It is frustrating. The data is there, but no one has bothered to write apps that make the comparison easy. For example, here is a census map centered on Seattle: Here is one for Vancouver: Both are good, but to compare the two is very difficult. It is apples and oranges.

    2. I’d imagine that other than Vancouver, the Canadian cities will be quite similar to American cities. Toronto, Montreal will be similar to large cities in the Northeast, and cities like Calgary are quite similar to the Midwest/Texas.

      1. My experience too, AJ. Every major Canadian city pretty much resembles the US city directly south of it. Reason pretty obvious. On North American continent, terrain forms like mountain ranges run north and south.

        Vancouver BC and Seattle are compressed by the western salt waters and the Cascades, also called the Sierras farther south. Same with the short-grass cattle country. And the farmlands, And the forests, as you go east.

        Clinching evidence is that Texas and Alberta both feature cities which have long had light rail.


  4. You’re my data crunching hero. I love this.

    When people talk about the Manhattanization of Seattle it’s helpful to remind them that we’d have to build 4 more Seattles on top of our Seattle to even get close.

    1. Since you’re an Engineer, Matt, I think you ought to know you’re talking Work in Progress.

      I met Architect Paolo Soleri at an exhibition in DC in 1971. His idea really was to create buildings which were themselves constructed out of giant cities.

      Probably merciful he never lived to take his luggage several miles on RapidRide to get it up to the Airport. Imagine that contractor doing four whole NYC’s worth of elevators.

      But talk about weighted density! Since one characteristic of beneficial density is extremely energetic activity and enterprise tightly compressed, would love to see Seattle locate the ST-10 LINK tunnel through Seattle to the Fourth Power at in the 300th floor level of Greatest! Seattle. Though we’d have to rename it IRT or BMT.

      We’d also have to call our subareas “Boroughs.”

      And just think of it! A deli (not the ones with bland ham and mayonnaise wrapped in plastic) at the head of every escalator, featuring corned beef sandwiches assembled from fresh rye bread, dill pickles, and of course Kosher Corned Beef purchased at huge Pike Place Markets located through all four of our New Yorks.

      No reason we can’t keep the wonderful Australian girl announcing closing elevator doors at Beacon Hill. But pretty sure that by now, we can match a borough- correct accent to every major station. No doubt about the one at 45th and the Ave! Also every curve can have actual rails tuned so train does “Rhapsody in Blue”.

      Riverton trackside neighborhoods will probably demand those noise barriers gone so they can hear every train turn loose the original opening trumpet call that Gershwin really did get from standing underneath a notorious curve.

      Newly-created Department of Continuity regs can help remedy elevator problems with system of tenement back-porches served by cast iron fire escapes. So that young passengers with switchblades and cigarettes rolled up in their T-shirt pockets don’t miss their trains to next Rumble at with the (Boeing) Jets at West Seattle Junction.

      Will definitely help getting in and out of UW and Capitol Hill LINK stations. And even better: Really do think it’d be a good promotion to have “Miss Proof of Payment.” New meaning to old form of dance. Bet passengers will break card readers screens off their posts as they go frantically clicketing past those yellow posts leaving a trail of chipped granite, like girls used to do with their “taps”.

      It’s our spiritual duty to fulfill Seattle’s original name in its First Nation language. Though pretty sure “Alki” can also be translated: “New York, But Don’t Hold Your Breath!” Anybody got a problem with that?


  5. I can see the utility of using weighted density. Still, there is a methodological problem. That problem is that some central cities have a larger proportion of a metro area population than others do. A place like Kansas City or Oklahoma City or Stockton covers developing areas that would not be part of the City of Seattle without massive annexation. Every state has different annexation laws, ranging from a “no annexation” structure of New England and New Jersey local government to “we can annex almost anything we want to serve” in a place like Texas. Consider that if Seattle extended out further, our density would be lower as suburban densities get blended into the average.

    It’s really a monumental piece of research. I’m impressed by that.

    With the data set already in place, I would think that a better measure that could be pulled would be to see how big of an area in each of these cities exceeds 10,000 persons per square mile, or how many people live in areas with densities over 10,000 persons per square mile. I realize that bigger metro areas would have bigger land areas or populations doing this, so maybe it could be tempered by a percentage of the metro area population or maybe tempered by the proportion of the central city land coverage or population.

    1. Yes, there is a methodological problem with just using density or weighted density measures. As an extreme example, slums and ghettos are very dense, but they aren’t very safe or walkable. So how do we build a sustainable urban core that also provides economic opportunity and stability for more than just the upper strata of society? The average nurse or public school teacher can’t afford to live anywhere near Seattle’s New Urban core. And when those essential members of a functional society are priced out of the neighborhood, what happens to the workers who don’t have college degrees?

      1. Well, we could try electing a Government that isn’t scared to call itself Liberal. Especially if they’re Democrats. And labor unions. And campaign finance laws. Worked before. Maybe just a coincidence, but also with a lot more transit.


    2. “Consider that if Seattle extended out further, our density would be lower as suburban densities get blended into the average.”

      That’s a dilemma with annexation. If Seattle had annexed Shoreline, White Center, Skyway, and maybe Burien in the 1950s, would we be better off or worse off? Would being in one city government have made them more urban, or would it have reduced the percentage of urban residents further so they’d be more outvoted? It partly depends on why suburbs developed the way they did. Annexation didn’t prevent northeast Seattle or far north Seattle from having twice as large lots and fewer corner stores than inner Seattle.

  6. In U.S., New Data Show Longer, More Sedentary Commutes
    From a transportation standpoint what matters the most to people is how long there commute time is rather than distance. Certainly sprawl tends to create longer commutes; witness the drive times into Seattle from Everett approaching the one hour mark. So I was a bit surprised to see that New York (the state) has the longest commute times in the US and the Dakotas are the lowest. There must be a “sweet spot” where density doesn’t reach the point of congestion. At the other end there must be a tipping point where it becomes faster to take transit (i.e. NYC) than to drive. With the possible exception of the downtown core, Seattle’s a long way from that point.

    1. Here’s some data by city:
      It’s interesting to compare Seattle and |Honolulu. Virtually identical commute time, Seattle has a slightly higher drive alone percentage and slightly higher transit usage. I would have guessed HI would have a higher walk/bike percentage than we do but the reverse is true. Where they make up the difference is in carpooling.

    2. It makes sense. Large cities like New York have a large number of workers. The city’s size means a long distance to cross, and a large number of workers from the suburbs means a long distance to pass their houses, so that’s two long distances combined. In contrast, when I went to Missoula for a friend’s wedding and somebody asked whether something was far from something, our host said, “Nothing is far from anything in Missoula.” Small towns have a short distance across and a small suburban ring. Some people drive sixty miles to another town for work or shopping but a lot of people don’t. And Montana and North Dakota have no big metropolises to drive to, so you don’t get e.g., people driving from Belfair to Seattle or Tacoma to work because there is no Seattle or Tacoma. People drive all over the state to see music bands wherever they’re playing, but that’s only every few months so it doesn’t add much to the total amount of driving.

      1. Nice try, but it takes 30 minutes to go just a few blocks in Manhattan because you have to walk to the subway station, go up/down stairs escalators to your track, wait for the train, and then repeat the process at the other end. Compare to Seattle: I can drive across the city from Madrona Park at Lake Washington to downtown in 15 minutes. If it’s a weekend, I might spend another 5-10 minutes finding parking and pay a few bucks if it’s Saturday and not Sunday. There is certainly such a thing as too much density, and NYC exemplifies it.

      2. it takes 30 minutes to go just a few blocks in Manhattan

        I’d think that if you only had a few blocks to work you’d walk? But NYC and Seattle are in a dead heat for “other” (i.e. not car/transit). I find that especially surprising since given the job density there is a much greater chance that you’re going to be living closer to where you work. Of course housing prices, which tend to correlate directly with density, probably mean that a majority of people can’t afford to live in the same borough as where they work?

      3. Do those commute times include everyone, or just New York residents?

        New York is dense, but really not that far away it drops down to single family housing and you are left with vast sprawl you need to commute through, just like anywhere else in the USA only with much larger distances involved.

        A commuter trip from Montauk to Penn Station is 116 miles and takes 3 hours on the Long Island Rail Road.

      4. Just a few blocks? In 30 minutes you should be able to go pretty much the entire length of Manhattan on most lines.

      5. To be fair, Kevin22’s concrete example was Madrona to downtown, which is about two miles. If we say from 34th to the library, that’s around 15 minutes on the 2 at noon, or 25 minutes on the 3. RapidRide G won’t quite go to Madrona but it’s supposed to be 10 minutes end to end. In Manhattan 15 minutes will take you from around the 10s to the 70s, or from the East Side to the nearest parts of Brooklyn and Queens.

      6. @Kevin22 — First of all, it doesn’t take a half hour to go a few blocks in Manhattan. It just doesn’t. People in New York City walk a lot and walk fast (faster and more than any other city). The subways are reasonably fast, frequent and not that hard to get to. As AJ said, you can get the entire length of Manhattan easily, as well as to and from Brooklyn, typically.

        One reason the numbers seem huge for New York is because of the big numbers of suburban riders who will put up with a very long commute, because it is fairly pleasant (drive a few minutes to the train, sit and read on it) and because there is very little alternative if you want that type of lifestyle (e. g. a nice house in New Jersey while being paid Manhattan wages). You also have the fact that way more people take public transportation to work than in other cites, and public transportation tends to be slower than driving. New York is no exception. When it isn’t, the drive is likely to be long.

        But again, it is ridiculous to claim that this means there is “too much density”. Just because you can drive quickly in a small city doesn’t mean that it is operating more efficiently. Far from it. Manhattan’s population goes from 1.6 million to 3.1 million during the day. To get that kind of increase in Seattle would mean way more people than can be found in Madrona. It would mean pulling in people from very distant areas. Remember the Seahawk parade? Remember how difficult it was to get all those people into Seattle, and how long it took? New York does that every day.

        As I said in the other comment, the advantage New York has over other cities is that it can enable more people to interact within a given amount of time. Measuring transit speed (or transportation in general) in terms of distance misses the point. You should measure in terms of people. In other words, people per hour, not mile per hour.

    3. The Dakotas case may be skewed by oil workers. As I understand it working on an oil field is like working on a ship. People fly in, work so many weeks nonstop, live in onsite housing built with the field, and then fly home for so many weeks off. So their daily commute is not very far, and it may be on company buses.

      1. Except the data in the top slide is from the 1990 census so the oil boom wasn’t yet a factor. The bottom graph shows commute times actually trending up sharply over the last few years in N. Dakota where the Bakken oil field is located and remaining flat for S. Dakota. I don’t know how they’d account for an oil field worker “commuting” from say Seattle but I’m guessing they’d take the travel time and average it over the period a person says on site.

      2. There is little to no “onsite housing built with the field” associated with the Bakken Field operations in ND. The operations that go along with fracking? You don’t want to live next door.

        Until the recent slowdown an apartment in Williston rented for Manhattan prices. So folks have been living all over north-central/northwestern ND and driving crazy distances to their field jobs.

    4. Type of travel matters, too. In cities with good commuter rail to downtown, officer workers will tolerate longer commutes because that time isn’t as ‘wasteful’ given you can work, eat, sleep, etc.

      In terms of transit ridership as a share, time competitiveness is only one part of the equation you also have to look at things financially, in terms of what Transit is substituting for. Studies have shown that one of the biggest drivers of Transit growth or decline is the cost of gas, i.e. cost of driving going up and down. Cost of parking is also a huge driver.

      People will take transit if it’s cheaper, even if it’s a much slower. This is why you see lots of poor people take the bus in cities like LA and New York, even it’s very slow financially it’s their best option.

      Of course, time competitive matters for the choice rider, which is the core audience of this blog, but that’s only one part of the population.

    5. NYC has the slowest average bus speeds in North America and likely the slowest average subway/metro speeds in the world. Public transit in NYC is PAINFULLY slow.

      1. Not surprising that NYC would have the slowest buses given the traffic volumes and a street grid that was laid out before the automobile. The subway is over 100 years old so there’s some inherent difficulties there as well. As for slowest buses in the world I’d take a guess that London has to be right up there even with their £11.50 congestion charge; I don’t think they let you jump on and off the back porch anymore while it’s “moving”. Seattle DT might be right up there by the time they finish kicking buses out of the bus tunnel.

        The point being, increasing weighted density brings with it acute transportation challenges that are expensive to mitigate. But we’re spreading out transit spending like peanut butter from Everett to Tacoma. And if you count the WSF system across the Sound and literally to islands of sprawl.

      2. They are slow in miles per hour, but not necessarily in people encountered per mile. I can drive across eastern Montana at really fast speeds, but my guess is I won’t find a Thai restaurant for miles. More to the point, public transportation is designed to move people to be closer to other people. In that regard, the New York City subway does really well.

        Stop spacing is also worth considering. While the New York subway has lots of stops that slow it down, there is a trade-off. It means that more people have a short walk to a train, which means that it may actually save people time. I have yet to see an analysis of New York’s subway which suggests that they added too many stops, probably because they didn’t. While this means that taking a train from Jamaica Station at the far end of Queens to Manhattan takes a while, it also means that folks along the way (in more densely populated areas) have a shorter overall trip.

        Of course in many instances there are express trains or buses that work better. These sorts of things make sense in areas like New York, because the density allows for it. I simply see no evidence (anecdotal or otherwise) to suggest that things get worse as a city gets more dense.

      3. >>Stop spacing is also worth considering. While the New York subway has lots of stops that slow it down, there is a trade-off.<<

        The NYC subway is much slower than it should be given it's stop spacing. There are two kinds of subway lines in the NY subway system:

        1. local
        2. express

        The "local" lines average about 1/2 mile between stops, and about 15 MPH. Toronto's Yonge/University and Bloor/Danforth lines also average about 1/2 mile between stops, but average about 20 MPH.

        My understanding is that there was some incident (maybe in the 1980s or earlier?, can't recall when exactly) where two trains collided in the NY subway that put future limits on their speeds in the name of safety. It may have just been as they approach stations. I can't recall the details – it was a few years ago that I read about it. The point is it's not just about stop spacing – the system is much slower than it should be given it's stop spacing. And the "PATH" system that connects Manhattan with NJ is only barely faster on average than the NYC subway.

      4. The ferries were originally for rural connectivity. I assume farmers from the west sound brought their goods to Pike Place Market the way farmers from Kent and Lynnwood did. That’s more “business” than “commuting”, and city-dwellers needed the rural food. Later a few people commuted to jobs in Seattle but most remained in the west sound full-time. The rise of exurban commuters in the 1990s is not really the ferries’ fault, and it’s also related to the collapse of jobs in the west sound (the former logging jobs).

        In Everett and Tacoma the sprawl was already there when the Sounder and Link projects started. The mistake was allowing the shortsighted sprawl development in the first place. Now two million people live in the north end and the south end, and we have to do something about transportation there. If we don’t they’ll have no alternative to cars and that will cause worse problems as well as stronger anti-transit sentiments. Also suburbanites are 4/5 of the population so they have the majority of votes so what they say goes. That’s why Sound Transit is so suburban dominated, because suburbanites are the majority and they aren’t “enlightened” about supporting a city-centric transit policy. The Link/Sounder extensions to Everett and Tacoma are no different in principle to New York’s PATH, New Jersey Transit, and Metro North trains, they’re just a lower (less comprehensive) level of service.

      5. I suspect the slow speeds of the NYC subways are due to old infrastructure, which cannot support higher speeds, and safety—there are other runs of your line running ahead of your line and you don’t want to rear end it, e.g the accident ChrisC mentioned.

        Express trains, depending on the line, quickly get you to particular stops and neighborhoods if you don’t want to slog through a multitude of stops on a local—-on the west side, you can take an IRT 2 or 3 instead of a local IRT 1; on the east side, you can take a 4 or a 5 instead of a 6.

        In spite of the slow NYC subway speeds, the subways, unlike link and metro, are ubiquitous all over the city and run 24 hours, making it very easy to live without a car.

  7. >> Nice try, but it takes 30 minutes to go just a few blocks in Manhattan

    I think you are missing the point of cities. Within 30 minutes of some spot in Manhattan, you can encounter literally millions of people. Within 30 minutes of any place in Seattle you will encounter fewer. In L. A. (unfortunately) it is the same story. The combination of density and good public transportation make it possible in NYC, but not in Seattle or L. A. (at least not yet).

    As Mike said, a lot of this transportation takes place on foot. Some of it via the elevator. Toronto, for example, could give New York City a run for its money, as an elevator ride plus a short walk (or subway ride) can allow you to cover a lot of ground. Well, not literally. Covering ground is not the point. Being able to meet people is.

    That being the case, there is no “sweet spot”. The denser the city, the less time it takes to meet other people. Even midsize cities that have little traffic to speak of don’t have the ability to interact with as many people as a crowded, congested city like NYC (or even L. A.).

    1. As Mike said, a lot of this transportation takes place on foot…. . Covering ground is not the point. Being able to meet people is.

      But the data by city shows New Yorkers aren’t walking more. And etiquette in elevators and on subways is to not interact. If you want interaction, take a cab! Which raises the question of do they classify taxi rides as transit or carpooling? In general I’ve found small town folk to be much more hospitable and interactive. Partly because if you said Hi to everyone you passed on the street in NYC you’d never get anywhere.

      1. What? I’m not talking about etiquette. I’m not talking about how many people you tip your hat to, or say “how do you do?” to. I am talking about meeting people. I am talking about what drives modern economies. Perhaps an example would be helpful.

        Let’s say that you want to meet with a hundred astrophysicists. For the sake of argument, let’s assume that one out of every 10,000 people are astrophysicists, and they are spread out evenly amongst the general population. That basically means you must draw from a million people. You decide to hold three meetings: One in Billings Montana, one in Manhattan and the last one in Seattle.

        It should be obvious to anyone that people spend way less time getting to Manhattan. Most of them can simply walk (and they do – I have no idea why you think people in New York don’t walk). Others take the subway. But you don’t even have to leave the island to get those 100 scientists together.

        For Billings it will take a while. If you are lucky, one or two can walk. Several have a short, pleasant drive. But most have to drive quite a ways, some from as far away as Boise.

        Now try the same thing with Seattle. Again, several people walk (from, say, Belltown). Several have a decent subway trip. But pretty soon you are trying to draw people in who have to drive or endure a tough bus trip.

        Now imagine you need 200 people, or that that only one in 5,000 people are astrophysicists, Cities like New York simply have a huge advantage over the other ones. There is no “sweet spot” — you simply want more density. This is why they build skyscrapers. It is why businesses are moving towards the central city (e. g. Amazon, Weyerhaeuser) not to the suburbs, as they did only a few years ago (e. g. Microsoft). You want to get people together, so you can do business. It isn’t just business, of course, but scientific collaboration, as well as entertainment. An obscure, specialized band can play New York City knowing enough people can make it to the show. In Billings they better have a following, because a lot of people won’t bother driving a half hour.

      2. Certainly easy to walk in NYC: a mostly horizontal city that enables you to walk very long distances w/ minimal problems, unlike cities with lots of hills, cough, cough. NYC’s density also allows for easier pedestrian access to locations for grocery shopping and other material needs.

      3. The citation was the second link I posted, <a href=""data by city. It’s from the Census and lumps walking, biking and work from home as “other”. NYC is at 14%, virtually in a tie with Seattle. Boston is the leader at 17.9%. San Francisco isn’t far behind at 16.5. It’s literally all over the map; no correlation with density, weather, hills or east coast/left coast. If I had to guess I’d say it might correlated closely with average age?

        As for your convention of astrophysicists, the top five convention cities in order are (weighted density rank); Las Vegas(+20), Chicago(8), Orlando(+20), New York City(1) and Anaheim(16). Again, no correlation with density, weather or walk-ability. It’s literally all over the map.

      4. Census data is only for commute to work, which according to the National Household Travel Survey is only 16% of overall trips made by a household annually. Ross’s data is for an entire day of travel, which includes non-work activities like errands and leisure. Unless you live atop a subway station or right next to a bus stop your “transit” commute is going to involve some walking. And let’s not forget that many transfers within the subway involve long walks between platforms.

      5. Census data is only for commute to work, which according to the National Household Travel Survey is only 16%

        The data puts the average amount of time spent in a vehicle at about an hour which means about 30% of that time is to/from work. And it aproaches 50% for the highest density areas. It’s still 83.4% of all trips by Private Vehicle, 14.2% Walk/Other and 1.9% of all trips by transit. The ratio doesn’t change much between work and non-work related which is a bit of a surprise since you’d think “choice” trips would tend to involve less driving. Then again, how many people have you seen drive somewhere to go jogging :-/

        Since 1990, the
        NHTS sample has only included households with telephones, so care should be taken in
        interpreting results that might be affected by telephone ownership (which is correlated with family

        This and other phrasing leads me to believe they are only calling land lines. If true, that seriously skews the data. Heck, there’s a growing number of people who don’t even know that you can “talk” on a phone :-0

      6. This and other phrasing leads me to believe they are only calling land lines.

        I took the NHTS last year and I don’t have a land line. My cell phone number is not even the local area code. I received a packet in the mail. These days you have the option to complete the survey online.

      7. Ah, I see that link was to the 2009 data and they appear to only publish every 6-8 years. So perhaps the next iteration (due out soon, I’d expect) will reflect a more modern demographic. It’s still decades ahead of what the Census website has based on 1990 data. But both seem to show that the more things change the more they stay the same. The only direct correlation seems to be that the more dense an area becomes the more time you spend traveling to/from where ever it is you want to be.

    2. Why walk when the subway runs every few minutes? But regardless of travel mode, density and mixed use means a lot of desirable destinations are within a mile radius or a half-hour transit ride, whichever metric you prefer. Running counter to this, a large city also means there’s another ton of desirable destinations outside this circle, so when people go to them it raises the average travel time. Equivalents to a lot of these destinations don’t exist in smaller cities so you can’t travel to them at all.

  8. Ranking cities in the manner is problematic. You immediately run into problems of designation, even when looking at population (do you use incorporated area, CBSA, or PSA?). In other words, where do draw the border?

    Trying to summarize density within a region is also very difficult. The population density approach is helpful, but I think it blurs the differences between cities. I think there is no alternative but to look at census data for each city to try and get a handle of things.

    To do that, I turn to this map ( Census blocks are not perfect, but they are way more granular than entire cities. The scales are wordy (e. g. 25,001 to 100,000 people) so I prefer a simply A to E scale (A being the most dense). That means:

    100,001 or more people — A
    25,001 to 100,000 people — B
    10,001 to 25,000 people — C
    1,001 to 10,000 people — D

    Below 1,000 is very low density. Typically this is industrial if it is close to the city, and agricultural or park land if it isn’t. With that in mind, here is how I look at some cities:

    Seattle — — We have only one ‘A’ block (in Belltown). Almost all of the ‘B’ blocks are within the city limits, within two big clusters. The main one is within a couple miles of downtown, while the second one is close to the UW. There are bigger clusters of ‘C’ blocks, especially in Seattle, north of the ship canal. There are some small ‘C’ clusters in Bellevue and south Everett, but most of the those ‘C’ blocks are found in Seattle. But even in Seattle there is a lot of ‘D’ mixed in with the ‘C’. Most of Magnolia, West Seattle, and the Northeast, along with parts of otherwise fairly dense areas (like Queen Anne and the C. D.). One of the more interesting things is how quickly density drops to the north. While there are smatterings of ‘C’ as you go south, there is not a single ‘C’ level block in Shoreline, Kenmore, Bothell or Woodinville. Of course this map was created from data that is already a bit out of date. Things have changed. But with the exception of Ballard and South Lake Union (which are both way more densely populated now) I don’t think there are any major changes.

    Phoenix — — By some measures, Phoenix has roughly the same number of people as Seattle. But the city is dramatically different. There are no ‘A’ blocks, nor even any ‘B’ blocks. It is a land of ‘D’, with a smattering of ‘C’. It reminds me of the quote attributed to Dorothy Parker about L. A., that it is “72 suburbs in search of a city”.

    L. A. — — The quote, of course, was from a long time ago (and many think it came long before her). L A. has changed. While it still sprawls, there is a center (of sorts) and huge swaths of ‘B’ mixed in with some ‘A’. At the same time, there are significant gaps as well as fairly distant clusters. Long Beach, for example, would constitute a major urban center in Washington State (second only to downtown) yet it is 20 miles from the center of town. Santa Anna would be the third biggest urban center if it was in Washington State, but it is 35 miles away. Serving an area like that with transit is a huge challenge. L. A. is a very big city both geographically and physically, with dense pockets everywhere.

    New York — — From a density standpoint, it is in an other league. Row after row of ‘A’ blocks in Manhattan, and quite a few in other boroughs as well. There is also a fairly solid transition from mostly A to mostly B to mostly C and so forth as you head east and north. New Jersey has a more “satellite city” look to it.

    San Fransisco — — San Fransisco itself has a strong cluster of ‘A’ blocks close to downtown. It also has a lot of ‘B’ and ‘C’, with very little ‘D’ within the city limits. Like Brooklyn, a lot of this is due to row houses, not apartments. Density never drops off, despite the low heights. Moving south, the border is essentially absent. Given the relatively small physical size of San Fransisco proper, this isn’t too surprising. Things then become very messy, as they do in L. A., but with bigger gaps between relatively dense locations. Unlike L. A. though, those areas are still not very dense compared to the center of town. Across the bay, though, there are parts of town that look more like San Fransisco. Plenty of ‘B’ and ‘C’ clusters that are fairly contiguous (although linear). To the north, however, density is extremely low. While it is common for people to say that San Fransisco proper needs to get more dense, I think the bigger problem is that surrounding areas (parts of East Bay and all of Marin County) need to add density. Like Shoreline, they are relatively close and sparsely populated.

    Boston — — Boston is, by various measures, roughly the same population as Seattle. But what is striking is how filled in it is. There are more ‘A’ blocks, but not so many that it is dramatically different. It is the contiguous nature and level of the density that seems so different. Draw around all the ‘A’, ‘B’ and ‘C’ blocks and you have a fairly large, fairly contiguous shape. It isn’t round, but there are few gaps within it. It is similar to say, the 5,000 foot contour around Rainier, which adapts for the rivers.

    Seattle just doesn’t have that. We have the big region in the middle, the UW, and lots of tiny lumps here and there. As Seattle evolves, it will probably have something like that extending to the north end (encompassing Ballard, Fremont, Wallingford, Greenwood and the UW). That might even extend in with a thin block to the northeast, to include Northgate and Lake City. Rainier Valley will also likely fill in. But even with all of that, it won’t be nearly as contiguous or as large as what Boston has unless there are major changes to zoning rules both in the city and the nearby suburbs.

    1. How many square miles is a census block? How does density per census block compare to density per square mile?

    2. Hopefully the next mayor will have the political courage to confront those single family neighborhoods north of the ship canal bridge and let them know that they will have to accept zoning changes to accommodate more density.

  9. This is a great analysis. There is a problem that should be pointed out though – census block (and census tract for that matter) data for the 2015 ACS is only available for 5-year estimates, not 1 or even 3 year estimates.

  10. Interesting though looks like Anchorage, AK (pop.: 300k) is missing (from the 250k+ map at least).

    That links to population weighted density by distance from City Hall for the metropolitan central city. So it will give you “concentric ring” data for Seattle, but not for Tacoma or Bellevue, since they’re subsumed into the Seattle data.

    Los Angeles is indeed a challenging city for transit. But you can see dense corridors heading west from Downtown Los Angeles to Santa Monica and (to a lesser extent) south from Long Beach. If you compare the city of San Francisco to central Los Angeles you find almost identical densities:

    I think the biggest problem in Los Angeles is the multiple employment centers and the weaker correlation of residence and employment location.

    As a former New Yorker, I’d challenge the idea that New Yorkers are unfriendly. I’ve found them more willing to help than residents of many cities. There’s just a tendency to be brusque, everybody thinks they need to be somewhere else an hour ago!

  12. More and more, I find the definition of “city” to be so variable that it isn’t useful. Take Indianapolis or Columbus, for instance. Those cities have instituted laws that effectively force annexation upon developers that wish to connect to city utilities, while lacking any of the long-term growth laws that exist in Washington or Oregon. As a result, Columbus continues to sprawl, with only a small number of incorporated suburbs, and Indianapolis has come to annex nearly the entire county! By comparison, cities like Seattle and Chicago are surrounded by dozens of suburbs. City of Medina? Town of Hunt’s Point? How did these become independent suburbs rather than a part of Bellevue? Accidental history, I suppose. As a “city” we should be looking at Seattle to include, at a minimum, Shoreline, Skyway, North Highline, and Burien, but probably a much larger area as well, since its effective reach extends well into SnoCo & South King, especially, if we are to compare it to some of the sprawling “annex anything that has a public utility” cities, otherwise, we are comparing apples and oranges.

    1. The tiny towns are old; they’re from the early 1900s. Beaux Arts was an artists’ colony. Hunt’s Point was settled in the 1950s, or at least that’s how old the house at the tip of the point is according to the Argosy Tour guide [1]. They were incorporated when tiny town incorporation was common, I guess. Bellevue was still a small town then too, with farmland around where 405 is now. The tiny towns are all along the lake shore because further east didn’t have as many people. There’s also Pacific, Milton, and Algona at the King/Pierce border; I don’t know their history but it’s probably similar. Or maybe that’s just how all towns were then and some grew or became industrialized and some didn’t.

      [1] The tour guide says that the people who own unchanged 1950s-70s houses along the shore are often out doing their own gardening or recreating in their yard, while those living in the McMansions with gondolas and professional gardeners never seem to use their yard, so why do they have it, and why do they bother with a $50K gondola that’s never used?

Comments are closed.