This is part 2 of a series.
In the previous post on this series, I looked at the idea of a residential land footprint – the amount of land each person takes up for their home. I also created a chart (updated here) that shows the different living conditions in the Seattle metropolitan area.
The best use of this land footprint curve is to have an easy way to compare the land use of our region to other regions. I chose five metropolitan areas that I thought would be interesting to compare to Seattle.
NYC
New York City stands alone in the US with regard to land footprint. The average person uses far fewer square feet of land. This is accomplished by stacking residences, allowing more people to share the same amount of land. Note that peak of 300 sf per person might be uncomfortable in a single-story world (keep in mind this includes your share of yard, street, and alley), but by building up each person can have as much floor space to live while using the same amount of land.
LA, SF
I was surprised by these two. I had imagined Los Angeles to be more sprawled the San Francisco, but they have very similar curves. It’s interesting to see where the curves cross – SF has more very dense construction, then the curves cross at around 500sf/person. LA has more moderate density units, and at around 4,000sf/person they cross again, leaving SF with more spread out housing than LA. Of course, LA’s metropolitan area only includes LA County and Orange County – and the sprawl I always think of extends beyond these areas.
Portland, Seattle, Phoenix
Portland and Seattle are similar. Seattle has more dense housing, Portland had more mid-density housing, and they align again as housing spreads out. However, the most common living condition is significantly more dense in Portland than Seattle (~3,000sf/person vs. ~4,000sf/person). I had included Phoenix just to show what I thought of as a large sprawling city, but was surprised by the curve. Yes, Seattle has many more people living 2,000sf/person than Phoenix, but Phoenix peaks near Portland at 3,000sf/person then stays lower than Seattle up through the higher footprint homes.
My Interpretation of the Curves (I’d love to hear yours in the comments)
Comparing Seattle to these five cities, it’s clear that we have a large land footprint. In the most common living condition, a Seattlite takes up a tenth of an acre for their share of their home alone. Add in the roads and freeways to get to your home (road area likely increases with land footprint), your children’s school footprints (there aren’t many multi-story schools around here), your workplace land footprint, the footprint of your grocery store, and we’re talking about a large amount of land per person. This is wasteful of our forests, our farms, and increases the distance and time needed to travel to any destination.
How do we move the peak of our footprint curve to the left? One way is through good transit. NYC started with an above ground railway over 140 years ago, which continued as a great subway system, and expanded from there. Building good transit that takes up little street space helps encourage development with small land footprints – your commute is shorter the closer you live to a transit stop, so building up allows for more people to live at these valuable locations. SF and LA had similar histories, except using streetcars and cable cars. Streetcars aren’t quite as good at concentrating residences as subways, as they stop more frequently. Seattle and Portland also had streetcar histories, but much of our growth came after streetcars were removed and cars became the standard mode of transportation. Of course there’s a lot more to land use than transit – for instance we’ll need to allow growth through zoning – but good transit is certainly a strong force in creating a small land footprint.
I think it might be a little more insightful and clear to have a Cumulative % chart, i.e. the same X-axis, but the Y-axis would be % of total population that have a residential land footprint < "x" square foot per person.
That way we can look at desirable density levels and quickly see how many people live within those density levels.
Yeah, that’d be cool if you wouldn’t mind adding that.
…on a separate graph of course.
Oh, and this awesome! Nicely done.
These curves are super-interesting, and I’d love to see them for more cities. Try some smaller metro areas, too, while you’re at it. (Spokane?) There are some small towns which I would expect to have bimodal density curves.
I agree that a cumulative “% of population with land footprint < x" may also be a useful way of presenting data and I'd like to see it.
This is really cool — great work!
A suggestion: try modeling a truly low-density, sprawling metropolitan region, just to see what the extremes are. (Since, as you found Phoenix is more medium density than low density, at least by US standards — probably because of water supply issues.)
How about Atlanta? Or Jacksonville?
The profile of sprawl has some surprises. People assume newest large city = worst sprawl but it doesn’t quite work that way. Desert cities like Phoenix have what’s essentially a hard urban growth boundary around their inner suburbs because of what you said, water. No growth unless more water is found, and only master-planned neighborhood developers are likely to apply, not people who want to build an isolated McMansion.
Seattle and the Bay Area are more in between, because it is possible to convert individual rural parcels on the fringe to McMansions and low-density housing, and eventually enough people do it that an exurb is recognizable. But the biggest sprawl is in the northeast because those exurbs don’t peter out like Marysville. Every fifty or hundred miles is another large city, so the exurbs merge and become megalolopis sprawl. Thus the paradox that New York’s metropolitan area is supposedly lower density than Los Angeles, because what people think of as New York is only 3 1/2 boroughs or so and then it drops off to a three-state suburbia.
“Thus the paradox that New York’s metropolitan area is supposedly lower density than Los Angeles”
No need for a “supposedly”. From my chart, NYC metropolitan area is far more dense than LA. It’s only similar to LA at and past 5,000 sf per person, which is a small percentage of blocks.
Careful about confusing metropolitan areas with proper cities.
Comparison of NY vs. LA densities:
http://www.uctc.net/access/37/access37_sprawl.shtml
Note: the curves above represent the metropolitan areas of these cities as definied by the US Census, not just the cities themselves. For instance, “Seattle” includes all of King, Pierce, and Snohomish counties.
[MikeP] Interesting. I wonder what their definition of NYC’s metropolitan area is – it has to be different from the Census definition.
Much as everyone here believes in Manifest Density I see an opposite trend..the dedensification of old cities.
I found this community in Phoenix..a brand new “agricultural village” at the edge or urbanity, called Agritopia:
http://agritopia.com/story-history/
Detroit is about to replace swaths of abandoned houses with timberland, in an interesting take on de-densification. It’ll be interesting to see how it works out. Sooner or later we’re going to have to learn how to shrink cities without them turning into cesspools.
I had thought that was the case too but a recent snowbird informed me that the majority of new subdivisions around Phoenix are being built on agricultural land. The residential use consumes less water than what was previously used for irrigation. So as more is “plowed under” the demand for water decreases.
@MikeP,
In order to have LA be more dense than New York, you have to include places that are so far from New York few people would confuse them with New York itself.
Plus, no one really trusts that study, anyway:
http://www.theatlanticcities.com/housing/2012/10/americas-truly-densest-metros/3450/
MikeP,
It’s more important to be careful not to assume that “central city” means anything at all.
If you study municipal boundaries, you quickly find out that they are extremely arbitrary. For example, look at Brookline, MA. North Brookline is indistinguishable from many parts of Boston, and indeed, it’s denser than places like West Roxbury or Roslindale (which are within the City of Boston). Or look at Memphis, TN, which includes a large amount of land that is practically rural.
To create a meaningful comparison between cities, you need to use a definition of city that is not so arbitrary. In practice, that means that you use a combination of density and distance.
We do need more urban agriculture and it’s not incompatible with density. We’ve zoned nature out of the city just like we’ve separated other uses. With our current technology there are plenty of opportunities for gardens on rooftops, gardens replacing lawns, pocket-park gardens, some goats and chickens, and reclaimed farmland in underused city and suburban parcels. The waterfront commission has some great ideas about mini-pocket nature reserves scattered around that would attract birds and critters back to downtown and help bring the ecosystem back toward fullness. If long-distance food transport gets in greater danger of being cut off, there will be increasing pressure to convert run-down/obsolete office parks back to farmland, and to revive more local agriculture along the exurban boundary.
I’ve often wondered about the data that goes into these comparisons. I assume that you are just using Seattle land area (~85 sm**2) as opposed to total city area (~140 sm**2), but I still wonder about the impact of other features like the proportion of unbuildable slopes and green areas, and the amount of industrial land and land set aside for port and shipping oriented activities.
Also, since Seattle has a bit of a reputation as a city with a high percentage of singles, wouldn’t a city with more singles tend to also show up as a city with lower density? I’m doing my part – soon to be married and cut my footprint in half without changing anything, but one has to wonder about the impact of singlehood on overall density.
That said, it would appear that Seattle has room to improve. And, with about 8000 apartment units coming on line this year alone, I bet we will be improving.
One of the benefits of this graph is it more or less ignores those unusable spaces. The areas used are broken up into “blocks”, which are actual city blocks in cities and chunks of land where there are no streets. If there are few or no people in one of those chunks of unbuildable land, it’s placed way out to the right somewhere. What I’m interested in is the left side of the chart, as the right side (beyond what I’ve charted above) is just farmland and parks.
As for singles/couples/families, I think you answered your own question. Yes, a city filled with couples living in the same sf/household as a city full of singles will look twice as dense. But it is twice as dense. Would starting up a marriage campaign complete with a publicly funded dating site reduce land footprints? I don’t know – I’m just displaying the data.
And congratulations.
OK, my bad. The data is granular since you aren’t calculating averages.
However I’m still curious about anything that could be skewing the data — topographic or otherwise. I’m always cognizant that this State is a bit different due to our *)(&^%!^ tax structure. This drives all sorts of odd characteristics in our business and reporting structures as people/corps try to work around things. Is it possible that something is being missed?
Maybe as a check, and out of a bit of curiosity, what does Spokane look like? And why? Spokane would theoretically show the same Washington specific effects that Seattle does, and topographically there are at least some similarities.
It really would have been more accurate to adjust for household size; which would be easy to do for city data. Seattle 2.05, Portland 2.27, San Francisco 2.30 and LA 2.82 people per household. But county data is “all over the map”; King 2.37, Pierce 2.59, 2.61. Marin County 2.34, San Mateo 2.74. Using MSA data and then labeling the graph and talking about as if it represents a city is very misleading. And using the MSA data doesn’t accurately reflect on a region. To compare “Seattle” with “San Francisco” you’d have to look at the entire bay area or CSA vs the Puget Sound. But then why not expand to a State level as you’ve got so many diverse jurisdictions you can’t really draw any conclusions.
The sf/person is fixed, no matter your household size – it would be the opposite of “accurate” to adjust for it.
I clearly described the data as representing the metropolitan areas in this piece as well as the others in this series.
I used the US Census boundaries for metropolitan areas. I’m not thrilled with them either, but wasn’t about to start defining my own boundaries. Besides, it’s less important at the left side of the graph which is what I really care about. Unless there are dense areas outside of the metropolitan boundaries, the left side won’t be affected.
OK, my bad. It doesn’t matter if there are four row homes with one person each on an 11,000 sq foot lot or a family of four in a 1/4 acre suburban home. The dots will be the same.
@lazarus:”with about 8000 apartment units coming on line this year alone, I bet we will be improving.”
No doubt, but as your marriage question points out, household size is a huge factor. My neighborhood in Cambridge, MA, has many more housing units per square foot than we had forty years ago, but we also have much less density. The apartment I live in with my wife and two kids was inhabited, forty years ago, by the Reilly family of NINE people–and that was the norm.
I think you should give some examples from the rest of the world, too. With a metric like density, USA is in its own dimension. What’s “dense” to us is hardly dense in Europe and esp Asia, and of course our “sprawl” is unmatches.
It would be great to get that perspective
I’d love to, but until the US Census expands its jurisdiction such a comparison will probably be too hard for a hobbiest demographer like me.
I’m sure we’ll have such data shortly, because the Census is obviously just a tool to implement the UN’s Agenda 21 plot to force everyone to live in Manhattan. ;)
The densest part of New York is pretty much as dense as anywhere, give or take 50%. It’s really everywhere else that isn’t.
I think people overstate how dense Europe is. Stockholm is about as dense as Seattle, so is Prague, so is Berlin (most of Germany, really). If you go to London and Paris and try to compare them to Seattle, you’ll find Seattle lacking in most regards. But comparing it to similar sized-metros and you’ll find the difference isn’t as stark as you might have thought.
Berlin comes in at about 10,000 persons / sq mi., which is pretty dense by most measures. I don’t know if that takes into account the significant acreage of lakes, parks, and forested areas within the city limits.
Eric H –
Seattle’s about 8,000 or so, so it’s not a big difference.
The Census Bureau has Seattle at 7250 per sq mi (2010) — 608,660 persons in 83.94 sq mi (land area). Berlin is 344 sq mi or so, which likely includes a fair amount of lake surface. But even leaving them in, Berlin’s density would be more than one-third greater.
Yes, that would still be in the same magnitude, more dense than Seattle, but still in the same magnitude. And this would be unsurprising, given that Berlin doesn’t appear to be a super-dense place. There’s some high-rise residential, lots of 4 and 5 story apartment buildings, and large helpings of single-family in the periphery.
So I’d agree that Berlin’s density is roughly comparable. However, the boundaries of Berlin extend deep into farms and forests, and include everything that’s remotely urban. Seattle’s boundaries contain a mere 17 percent of the metro population.
Or, put another way, both greater Seattle and greater Berlin are home to about 3.5 million people. Berlin squeezes them into about 350 square miles. Seattle puts 3 million into a Census-defined urbanized area of 1,010 square mi, with another half million sprinkled into the non-UZA parts of the metro counties.
So with Berlin packing about as many people as Seattle into about one-third of the area, I’d consider that as adding up to more density for Berlin. Your measurement may vary, of course.
Paris within its city limits has a density of 54,900/sq mi. This compares favorably to NYC (27,013/sq mi) and even to Manhattan (69,771/sq mi).
Paris has just one skyscraper within city limits (they decided it was too ugly and outlawed building new ones). But the other buildings, and I mean ALL the other buildings, are 5-8 stories high. Proving that density it not necessarily a consequence of building height. Building every building moderately high (as is typical all over Europe) creates more density than building a few buildings extremely tall (as is typical in the US).
This is very informative. Thanks!
I love it! This is right up my alley, so to speak, and I’m going to be looking back at this in the future.
I think the secret to moving part of the graph to hump up on the left is not just density of people, but density of use. Seattle’s getting better at the former, in the city at least (though I think we’re losing the war out in the far suburbs), but we’re not getting better at the latter. Which might be why we’re losing the war.
Density of use means having stuff close enough to walk to. Like “Walkability Score”, if Walkability Score wasn’t full to the brim with garbage data (my house is listed as “60”, when it’s realistically more like “15” — no, the place that makes Japanese pastries for the restaurant trade is not a “grocery”. Density of use is more than “is there something that might pass as a restaurant nearby?” but “is there a community that meets the typical range of everyday needs nearby?”. Which means BOTH a full-service supermarket and a quick-stop convenience store a block away, and a dry-cleaners and a liquor store and a newsstand (yeah, right) and multiple restaurants in both the $5 and $20 categories and the usual assortment of other needs.
Because building lots more apartments and condos moves the graph, but if they are filled with check-cashing places and real estate offices and nail salons (as valuable as nail salons are to the immigrant community), you don’t have a real neighborhood there. And it is virtually impossible to build one from scratch; the functional neighborhoods I mean are the same ones that were here in the 1920s. The closest thing to new ones are not walkable but drivable, like Bellevue’s immigrant-populated strip malls (themselves a 1960s concept, not a 2010s one).
Seattle is not losing any wars.
And your comment about the immigrant community and nail salons is vaguely offensive.
Nail salons are a popular way for Asian immigrants to get a foothold in the economy. What is offensive about that?
I agree that the nail-salon comment is offense.
Again, explain to me how it is “offense”. It is not remotely offensive. The only way I can even imagine that someone would find it offensive is if that person found immigrants offensive, which I patently do not. Indeed, I find immigrants in general to be markedly less offensive than native-born people — and immigrant neighborhoods to be far more interesting and affirming than white neighborhoods.
But since two people now think I’ve committed some kind of a hate crime, I’d like to hear why you think that.
It should be obvious to you why this is offensive.
Well, it’s not. You’re going to have to help me out, sunshine. Educate me.
The vibrancy of our urban environment is not coming from $150-a-cut hair salons or Restoration Hardwares or rows of frat-boy pukeatoriums, but from the people bringing their customs and products and ideas here from elsewhere. The fact that these people are priced out of many neighborhoods is to the detriment of those neighborhoods, not the poorer, underserved, undertransited ones they are economically relegated to. The sterilization of wealth and inequality is destroying large swathes of this country, including here in Seattle. Now THAT’s offensive.
Don’t be cheeky.
I have no idea what you are talking about with pukeatoriums. And I don’t know why you think immigrants aren’t be allowed to participate in those. I don’t think “These people” aren’t priced out, I know a number of immigrant friends who live places like Kirkland, Belltown, Capitol Hill.
I know for a literal fact that saying that white people suck and immigrants are awesome is racist, which is by definition offensive. Also a lot of white people are immigrants (this one is), so there you go.
“Pukeatoriums” was a dig at Fremont, which is a bit that way on Friday and Saturday nights.
“I know an immigrant who lives in Belltown” isn’t enough to counter the plain fact that FOR THE MOST PART immigrants are people of color and FOR THE MOST PART they not Microsoftie H-1Bs and FOR THE MOST PART they and other people of color, who are traditionally dealt with as if they were immigrants in Seattle, are being driven out of the city or not allowed in in the first place, and are taking up large areas of the south county and beyond (and other places in the outer ring too). There are exceptions: Northgate has a lot of immigrants. But Northgate is in NE Seattle, which was recently ranked as the fifth-hottest real estate market in the country, with prices going through the roof, which means poor people won’t be living there much longer.
Neighborhoods that poor people can’t live in are sterile. I didn’t say or imply that “white people suck”, but I did imply that well-off white Seattleites are not a people full of cultural vitality in their own right. Which is true. They’re full of money, though.
The median listing price for a home in Kirkland or Capitol Hill is $600k; in Belltown it’s $625k. The people working in the nail salon or the White Center taqueria can’t afford that.
I get it. Fremont does suck. You’ve got an incredibly over-simplified view of immigration and vitality, but go ahead and live in White Centre if you want.
Median value of owner-occupied housing units in Kirkland according to the last US Census is $478,800. I don’t know where you made up the number for Belltown but it’s nothing new for people working in the neighborhood to not be able to afford to live there. Teachers on Poverty Rock can’t afford the median $893,700 price of owner-occupied housing but it still has one of the best public school districts in the State.
One of the reasons I live in the Des Moines Marina District is it’s surprisingly walkable. There’s multiple restaurants within a few blocks of my apartment, a convenience store on the same block, and a grocery store five blocks away. People think of Des Moines as a seedy, sprawling place because that’s what it’s like along 99, but it has a very walkable downtown.
The unfortunate bit is the most walkable parts of Des Moines are the parts that are worst served by transit. There’s Rapid Ride along 99, but Marine View drive is only sparsely served, mostly by the 121/122.
The 166 hits almost the full length of Marine View Drive, and runs every 30 minutes all day…
Yes, but it’s an extremely slow way to get anywhere. By the schedule it takes 30 minutes to get to Kent, or 20 minutes to get to the Burien transit hub, and it’s usually slower than that. I stopped trying to use it to get from the Kent Des Moines park and ride to home because it added three quarters of an hour to my commute — a 30 minute wait (because the 197 usually just barely misses it) and then a 15 minute ride while it went clear down to Reith Road before doubling back. Not worth it when I can drive home from the park and ride in ten minutes. Most of that is spent waiting for the incredibly slow traffic lights at Military and SR-99.
The eventual Highline CC Link stop should make the 166 work for you… in many, many years. Sigh.
I’ve never been to Des Moines. I should check it out. If I had my druthers, and could stand to live so far from work, I’d probably move to downtown Burien, which has most of what I need right in the village (albeit some of it behind ten-acre parking lots still).
I live in a “immigrant populated” neighborhood. What I see is that for some, walking can mean going a half mile or even a mile to store and back. And is that wrong? I mean what is the point of calling something walkable when all you is get on an elevator, go downstairs, walk half a block to your coffeehouse and sit back down and web surf on a Kindle? My family grew up in Queens walking everywhere (or biking) because my Mom who was at home all day didn’t drive, so the experience here is familiar to me.
What’s wrong is that the supermarket is not closer, that you have to walk through a large parking lot to get to the front door, and that the other large parking lots around exacerbate the distance you have to walk. Was your Queens neighborhood really like East Hill? Do you really have as many choices in a 5-minute, 10-minute, 20-minute walk in East Hill as you had in Queens? Do you have to drive or take a bus out of the neighborhood more often in East Hill than in Queens? Which part of Pugetopolis most resembles your Queens neighborhood?
Again you’re trying to twist words to mean their opposite. The whole reason the word “walkability” exists is that people want to walk 5 minutes to a coffeeshop or supermarket, and think that building unwalkability is unnatural and harmful. Twisting the phrase to say a mile walk to a strip mall is “walkable” and all these coffee-shop denizens just don’t realize it, is wrong. The isolated cul-de-sac land I grew up in was walkable all along, I just didn’t realize it! Silly me, I should have just enjoyed the mile+ walk past single-family houses to the supermarket, and the hourly bus I thought was crap was actually wonderful. But I did consider myself lucky that I had an hourly bus within five minutes’ walk, because most of my friends didn’t.
Again you’re trying to twist words to mean their opposite.
Once you say that, there’s no reason to have a civil discussion with you then.
If you ignore a simple fact and walk away as if it were the main point, it shows you’re not serious and are in denial. I could mention that’s a pattern among Republicans recently. You could dispute this fact directly and say it’s not true. Or, much more productively, you could address my first paragraph.
Gotta stick up for “John B my grandfather and me” here. Well, more like my grandmother since she’s the one that did all the walking while grandpa took the tube in to Barclays Bank. Americans in general have to get past the idea that the only place you traverse more than a 1/4 mile is in the gym on a treadmill. I “got to” walk home from S. Kirkland P&R yesterday; a nice 45 minute jaunt since OBA told me I’d missed the 249 by 1 minute and the next one was 49 minutes away (plus a 10 minute walk). The “instant now” idea is a perversion of what “walkable” used to be. It’s become the false ideal that transit + zoning can miraculously duplicate the product of the automobile society.
no, the place that makes Japanese pastries for the restaurant trade is not a “grocery”
Tokara is definitely not a grocer, but her wares are delicious, and you should certainly take advantage of your walkable proximity on the 3rd Sunday of the month! (Place an order the day before via voicemail.)
While I agree that Walk Score’s binary algorithm (“Is or is there not one of [x] within distance [y]?”) is fundamentally flawed, I have to imagine it was counting the (defunct) Phinney Market as your actual walkable grocer.
Nope, it was counting, first, an “Ethiopian Grocery” in a house a couple of blocks away, which from its street address (there is no “Park Avenue” in Seattle) I was able to figure out was actually in Renton, and then Tokara, and then Phinney Market (which is not a market anymore) and then the imported food shop across Phinney — but not 7-11, which is weird. And of course none of those are full-service groceries. And even the limited ones are up a steep hill.
For restaurants, it claimed that there was a Subway sandwich shop right on the corner of 60th and Aurora, which is ludicrously wrong (the address in question turned out to be in Kent, not far from Mr. Bailo).
There were other errors as well. My hobby is correcting bad markers in Google Maps, particularly in Mexico, and I used to try to fiddle with Walkscore too, but I ain’t got time for that.
Wow.
I had just been thinking that my neighborhood didn’t have enough places to grab Cigarettes, Spices, Beer & Injera. Shame that you were so close to having such a thing, and yet so far.
Does Ken’s count as “full-service”? It’s about as big as any medium-density area could conceivably support, especially without dedicated parking.
I would count Ken’s. They’re small, but they have a large fresh produce section and a deli counter. They’re too far to walk, though; not to zone you directly into my house, but they’re almost precisely the same distance away as the PCC on Aurora, which is flatter. I’ll cycle to PCC but not walk, and I won’t do either to Ken’s unless I’ve got a bunch of other things to do up that way. That hill is a mother.
If I was designing a walk score algorithm, the needle wouldn’t start to move until you hit a certain density of uses — not just one store but two, and three or five retail shops (minimum one bookstore) and three or five restaurants, etc.
PS — your Streetview captures my favorite sign in Seattle, now sadly gone — the small one to the left with the picture of a pizza slice, hot dog and sandwich and the word “PANINI’S”. Ah, we don’t appreciate our greengrocer’s apostrophes until they are gone.
Thanks to Amazon and ebooks you’re going to be revising just about everywhere down to zero unless there is a public library in the vicinity. And the extintion of the neighborhood grocer will be complete if Bezos or his competition can crack that nut.
I’ve been known to barrel my way up 64th from Woodland to Phinney on days when the 5 was FUBAR and the 358 was a better option. It’s surprising close and quick once you get over the steepness. I just lean into and move quickly and suddenly, miraculously, I’m on the crest.
That said, I wouldn’t want to do it on a bike, or every single day, or for every possible tiny errand.
Walkscore says its algorithm is limited and its results are sometimes off. It doesn’t distinguish between comprehensive supermarkets vs limited convenience stores, and it doesn’t understand how hills and lakes impede walking. But it’s valuable as a rough guide of which neighborhoods to consider further and which not to, and the real estate industry with its “I have a walk score of 79!” is introducing it to people who don’t know about it or hadn’t thought about walkability before. That will hopefully bring the issue of walkability further into the mainstream of society.
The reason nail salons are noticeable is the lack of a diverse variety of businesses around them. But remember that these areas are recovering from redlining. The Vietnamese and others who first brought businesses back to MLK and east Jackson started with one-room restaurants but have gotten into small strip malls with a wider variety of businesses (even if they mostly cater to the specific needs of an immigrant community), but the next generation of those will hopefully be larger and denser and include more businesses addressing the needs of the wider non-immigrant community.
Even by the standards of “limited”, the algorithm is limited. It’s almost as if they don’t get that knowing the density of various types of activity within a given radius is far more telling than just knowing whether or not one something exists.
If it’s so easy to make a better Walkscore, why hasn’t anybody done it yet? Hobby opportunity, here.
I had a lovely experience with Walkscore and Google Maps trying to find a hotel with the best walkability and transit in the Santa Clara wasteland. I searched along the light rail line and follwed up on names like “Ralph’s Store” or something. Google Street View was the most illuminating; it showed that me Ralph’s Store was a tiny convenience store that probably closed at 5pm, and that the other store had only Indian goods. I ended up in a hotel on Mission View Drive, with an hourly bus. It was the only place cheap enough for my company to consent to. I limited my trip to the shortest possible time, to get out of Silicon Valley as quickly as possible.
Kirkland isn’t recovering from redlining and it can hardly be called enthnically diverse (White persons, 79.3%) yet there seems to be more salons between Totem Lake and DT Kirkland than any other business. And what’s strange is they are virtually all located in new construction. I guess that’s what you get when you build more “mixed use” than you have use for.
No, that’s what you get when you build “mixed use” that’s 80% parking infrastructure at ground level, where the retail “front” is as false as that word can imply.
Nail salons just happen to be a low-margin business that appears anywhere commercial storefront property is cheap. Check-cashing stores are another one. In the 90s, the classic example was the beeper store. (Rember beepers?)
DP, what are you calling false mixed use?
Oh, look: a tanning salon, a (contract) Sprint store, a dry cleaner (work performed off-site)… What all these have in common is a total lack of depth or customizable space, little-to-no walk-in appeal, and doors locked by 5:30 pm.
Why are these things invariably all you find in buildings like this? Because 80% of the ground floor behind them is parking.
That’s false mixed-use. Nothing mixed about it.
Bless you, d.p.
I think the reason you were surprised by the curves is that they don’t really paint much of a picture of the built environment. QuickFacts has a line for percentage of housing units in multi-unit structures. Seattle and LA are close at 50.4% and 53.9% respectively. They’re also similar in density at 7,400 and 8,100 people per square mile. SF has 67.3% multi-unit structures and a density of 17,200. Los Angeles is as you imagined far more sprawled than San Francisco. Portland is more like Tacoma than Seattle with only 37.9% multi-unit housing and a density of 4,375 people per square mile (Tacoma; 34.9% and 4,000).
What percent of the total city population is shown in the graph above. I’m having a hard time connecting, especially since Seattle/Portland/Phoenix never exceed 1.5% or so in the graph.
The percentages are for each group of 100 units on the bottom scale. For example, in Portland it looks like 1.8% of people live between 3,000 sf/person and 3,100 sf/person.
The area under each curve has an area of 1. So if you look Portland, around half of the area under the curve lies to the left of around 3,700*, which would be the median living condition.
* If you assume the lines continue downward fairly quickly after 8,000, which they all do.
2000 sq ft per person…by my standards you guys are Land Hogs!! Here on Kent East Hill, I live in a 850 sq ft apartment (which is unusual to begin with, since many are immigrants with higher than average family sizes). But then the building is stacked 3 stories. Assuming a least dense case of 1 person per all 3 apartments, I’m at 280 sq ft, or nearly an order of magnitude less than “urban” Seattle!
Not unless you access everything via communal hallways. You have to add on the square feet of parking, landscape, vacant lots, etc. to get the land footprint. A one room cabin on 40 acres does not a dense land footprint make.
So Seattle doesn’t have streets, parking, garages and lots?
And do you factor in Greenlake, and Woodland Zoo and the other massive parks? University of Washington?
Seattle buildings have parking, etc., but not like apartment complexes in the suburbs do. The complex I used to live in in Kent had far more space devoted to parking lots than to buildings, and featured acres of manicured lawn that wasn’t useful for anything.
I think the answer is “yes, but Kent East Hill has more – at least when you count the entire area.” (And didn’t someone else say that these statistics are averages for all the Seattle metro area?)
Residential land footprint is calculated by taking the land area of a block (to the centerline of all streets) and dividing by the number of people living on that block. So it does include Seattle’s streets, parking, garages, and yards.
Because the measurement is done at the block level, uninhabited areas are far off to the right and not interesting for this comparison. I’ve called it “residential” land footprint because we’re looking at living patterns. I suppose similar curves could be built for an employment land footprint, retail land footprint, recreation land footprint, etc.
Also, it’s the metropolitan area – so Kent is in there as well.
Is the land footprint of that cabin really 40 acres? Is the answer different if the residents are self-sufficent and use all of that land to supply their needs?
In the urban context, it’s a little strange to say someone’s land footprint is only due to their residence and some immediately surrounding infrastructure.
Actually the cabin on 40 acres only has a land footprint of 2.5 acres. Because of the “block” methodology no one can have a footprint bigger than 1 block. It also has the effect of removing large parks and penalizing “pocket parks”. Conventional suburban business parks would also be “out in the weeds” and not skew the land footprint.
I would say whether they’re self-sufficient or not doesn’t really figure into the answer. The fact is we have a world population of nearly 7 billion and it’s still growing. Even if people were interested in reverting to a subsistance-farmer lifestyle we couldn’t give them all 40 acres to do it on.
Mind you, once we run out of fossil fuels I don’t think we’ll have much choice; the fact is dense cities rely on very energy-intensive transportation to feed all those people and haul away their garbage. That’s probably a good 100 years in the future, though.
Bernie: you’re wrong about some of that, because the 40 acres is usually listed as one “block”, and the same with the entire suburban office campus. (You’re right about the large parks.) Block sizes are larger as you go out further.
Now, if you’re out so far that there’s a “back 40” with no road access, that *does* get deleted from the measurements.
“dense cities rely on very energy-intensive transportation to feed all those people and haul away their garbage”
London and Glasgow and other cities were 2-4 story medium density for hundreds of years before gasoline. Highrise towers may not be sustainable without long-distance food transport, but the sustainable level is significantly denser than most American suburbs or the large-lot third of Seattle. And even with those towers, we haven’t even begun to tap the possibilities of solar power or rooftop/indoor gardens to make them more self-sustaining.
I’d like Matt to expound on that. He said:
Now, if you’re Willy Nilly changing your unit of “block”, titling the post as “Seattle is Not Dense” and labeling graph lines as Seattle and SF to represent Seattle-Tacoma-Everett and San Francisco-Oakland-Fremont [ad hom]
Horse pucky as Colonel Potter might say. Try to duplicate London today with horse power and it would be in deep shit.
Bernie: you’re wrong again; the statement about London is roughly correct.
Mind you, “before gasoline” isn’t as long ago as you’d think: London was powered by coal and steam, not just by horses.
A measurement of LA without parts of San Bernadino isn’t really a measure of LA.
But it is, for some purposes. The statement that “New York is less dense than LA” is useless to me because I only care about the high- and medium-density part of New York. The lower-density suburbanish parts should be counted separately that’s a whole different issue and attracts different people. What I want to know is, “How does the higher-density part of LA compare to the higher-density part of New York?” “How much of LA is equivalent to the least-dense San Francisco neighborhood?” How does the row-house-or-denser part of Seattle compare to its counterpart in Portland, LA, and other cities? What percentage of the city or metropolitan area does the higher-density portion cover? Those are the kinds of questions that matter for me. As for San Bernardino, the first thing I care about is how frequent is the transit between LA and San Bernie, and does it run weekends? (I looked at this for Chino Hills. There seems to be a nice commuter rail to Ontario, weekdays only. There are probably full-time buses to Ontario but Chino Hills seemed more iffy.)
Yes, “New York is less dense than LA” is stupid.
Imagine two cities: Crowdedton and Sprawlsville.
Crowdedton has an island that’s 10 square miles and has 1 million people, and a bedroom suburb that’s 100,000 square miles and has only 10 people (this is riduculous, obviously, but it illustrates why that comparison is stupid). Average density is thus, 10 people per square mile.
Sprawlsville is pretty much the same density everywhere, and its 1 million people stretch over 1000 square miles, so it’s 1000 people per square mile.
Now, the vast majority of people in Crowdedton live in a very dense community, and the average person in Sprawlsville lives in sprawl, so the average density isn’t really a very useful metric. That’s pretty much the case with New York, most people live in pretty density communities, but the spread out part is far more spread out than it is in LA. But really few people live there, so the metric isn’t actually that useful.
San Bernardino and Riverside are as much a part of the LA metro area as Bridgeport, Trenton, and the Poconos have to LA and OC proper. There is now seven day a week commuter rail service to San Bernardino and Riverside County, and both counties have comprehensive suburban bus service, some even running every 15 minutes, 359 days out of the year.
The LA County Congestion Management Program factors show that, quite contrary to popular belief, most people commute to places in their same area or adjacent to it. The super commuter from Riverside to Downtown LA or South Coast Metro is a small, small minority, yet it brings up a lot of attention, just like the person waking up at 4 am from Pennsylvania to get on a bus to Midtown Manhattan. http://www.metro.net/projects_studies/cmp/images/CMP_Final_2010.pdf
I think these graphs are really cool… here are some reasons I can think of for differences between perception of these cities and their curves:
1. What’s included and what’s not. People are surprised at what the LA area includes and what the SF area does not, for example.
2. Blocks with zero population aren’t shown at all. Such blocks include freeways and interchanges, parking lots, rail embankments, parks, office blocks, industrial areas, big-box stores. All these represent things necessary for cities (transportation, recreation, retail, commerce, industry). In cities where uses are closely mixed these uses dilute the appearance of residential density even as they contribute to walkability; in cities where uses are separated they can make residential blocks appear quite dense while increasing walking distances. For example, in SF’s Sunset District the L train runs down Taraval, which has lots of ground-floor retail and professional services with apartments above. My friends that live on Taraval do most of their shopping and transportation on blocks with residences on them, and work in an area with mixed-use blocks as well; even some nearby parks are located such that they might be counted as parts of residential blocks. When I lived in Santa Clara my apartment complex had the block to itself; my shopping, workplace, recreation and a good deal of my transportation took place on blocks with no residences on them whatsoever. All these things were necessary for my life but wouldn’t have been counted toward land footprint on this graph at all, where for my friends in the Sunset they would be. This is where aggregate density measures can actually be more telling.
3. Commercial density isn’t represented at all; I suspect a big part of the perception of some cities’ density has as much to do with commercial density as residential.
4. In some cities well-known attractions and famous elite residences are concentrated in dense places. You could very well vacation in Chicago or Seattle or Boston or NY or SF without ever getting far from downtown. You could very well vacation in LA or Vegas without ever getting near downtown. In both cases you’d be missing a lot, but it’s a real phenomenon that shapes the way we think about these cities.
(On #2, even light industrial buildings share blocks with residences in many parts of Chicago and New York at least… I’ve never been to New York but it’s mentioned in Death and Life. This has diminished, of course, as industry has consolidated, moved out of dense cities, and moved out of the US.)
#2 and #3 are important points which distort the analysis.
Matt,
Do you have similar data for the area inside just the city limits?
Thanks….
Yes. Check out this discussion about how to pull the data yourself, if you’d like. The data comes as a row for each block, with identifiers for what county, city, and even exact longitude/latitude for those blocks. You can grab only rows for the city you want to look at, and with a bit of effort even a piece of a city (in Excel, exclude rows that are past a certain longitude/latitude).
However, I’m not too interested in city limits. They’re arbritrary political boundaries, and might cut off the dense areas I’m interested in for some cities, while not for others. I find by taking the entire metropolitan area, this kind of cutting-off is done on the right side of the chart above, which is less interesting and more predictable anyway (we all know what sprawled homes look like).
I disagree with respect to the usefulness of city boundaries. It’s by and large up to city governments to enact zoning, build infrastructure and promote business that shapes density. And I find it unlikely that a significant pocket of highrise housing is going to get left out. What drives vertical development is sky high land prices which are going to be concentrated around the CBD. Regardless, the percentage of cliff dwellers doesn’t really say much about overall land use. Take UW for example. It’s roughly 600 acres and there’s approximately 5,000 residents. Blocks of residence halls would make it look like NYC but averaged out it’s more like 5,000 sq-ft per person. Miles of auto oriented big box stores and office parks as you’d find outside of the DT core in LA define the city much more than it’s small percentage that live in tall buildings.
If you cut the chart by city boundary, the charts would change shape drastically based on the way the city boundary is drawn, which often has little to do with how people live. Incorporated Jacksonville not only contains all the metro’s land, but nearly all its people; incorporated Miami holds very little. LA, Houston, San Jose, and Seattle have all annexed suburbs or unincorporated areas after they were mostly built out; San Francisco hasn’t to such a degree (I think incorporation preceded development in most of SF).
The difference to a chart like this is less dramatic than the difference in aggregate density numbers, but it’s significant. Consider Seattle, whose population is less than a quarter of the metro’s. Most of the dense blocks are in Seattle, so the left half of the graph jumps up quite a bit as dense block count shrinks slower than the overall denominator.
In fact, when metro area boundaries have to be cut at lines across which large numbers of people regularly commute, the same sorts of weird effects come into play. LA and SF are part of megapolises that the metros used here only partially cover.
San Francisco actually is a consolidated city and county. In other words, you could consider all of San Francisco unincorporated area (as the Thomas Guide did when it colored it yellow). The State consolidated the two city and county governments and sent the rest of San Francisco County to create San Mateo County.
San Francisco is a lot older than LA. LA didn’t reach 100,000 people until the 20th century, while San Francisco was there around the time of the civil war. On of the main reasons Los Angeles got big was the 1906 earthquake that convinced a lot of people San Francisco might not be such a great place to live. So Los Angeles really boomed at the same time as the automobile, which is why it’s always been so sprawled out.
Los Angeles really was all about water. But yes, a generation coming back from WWII and the rise of the automobile defined what LA looks like today.
And LA didn’t REALLY boom until the postwar defense and aerospace explosion, when cars were part of the mythos of the city.
Funny, I was just reading about kids in South Central who couldn’t care less about cars — they can’t afford gas — but are all into ultra-customizing their bicycles like lowriders used to with 1962 Impalas, and riding the Red and Blue Lines to Boyle Heights where the entertainment options are.
Where were you reading about those bikes? I would love to check that out.
http://la.streetsblog.org/2013/01/22/black-kids-on-bikes-spread-the-gospel-of-the-bike-at-the-king-day-parade/#more-80016
The Bay Area is far more dense than Seattle Metro any way you slice it. There are suburbs on the Penninsula (San Mateo, Burlingame, etc) that have corridors that are denser than much of Seattle. And the East Bay Combo of Oakland/Berkeley is similar density to Seattle with only a slightly smaller population. Of course, San Francisco proper absolutely blows Seattle proper out of the water in terms of density (literally, more than twice as dense, with very few detached single family homes, unlike Seattle).
According to the Census “only” 67% of the housing in SF is multi-unit. However, I suspect that the homes shown in this picture are not classified as multi-unit although I don’t think they can really be considered detached and as the story points out many aren’t truly single family anymore.
Good point. When you have a popular and dense downtown with a very restricted housing supply, the illegal sublet starts happening frequently.
Those rowhouses in SF are not Single Family Homes in any, way, shape or form. They have no yards and the indoor square footage is usually twice as much as the lot size. There are corner stores on most blocks in those neighborhoods.
SF is not just dense downtown, it’s dense the entire city. Remember SF has nearly 850,000 people in about 46 square miles (17K per sq.mile). Seattle has 620,000 people in 84 square miles (8K per sq. mile).
It’s obvious when you walk around SF – the entire city is “city” and urban. Even the residential parts are rowhouses with corner stores, bars, restaurants, etc. Outside downtown, Seattle has patches of urbanity in sea of suburban-like single family homes. The compariosn isn’t even close.
That doesn’t mean they’re not Single Family Homes. I’m pretty sure they were originally built as single family homes and built with pretty much the same “yard” distribution as they have today. When you convert a laundry room to a rental unit then yes it’s not single family any more.
Lotsa SF homes in Seattle proper. But the Density is coming. Big time. 1964 to 2013, the trend cannot be denied. Density is alive and well and rapidly expanding.
I’ve come to the conclusion that not only is this posting wrong, but it is actually dangerous to the cause of transit in Seattle.
Why? Because the title clearly states that “Seattle is not dense”. While the data presented might be accurate for the metropolitan area, the title surely is totally inaccurate.
Al Diamond says it best in his post about city limits. If this data was presented just for Seattle city limits, it would show much higher density and a completely different distribution of densities. Basically the density story inside the Seattle city limits is totally different than the density inside the metropolitan area.
Why does it matter? Because if the takeaway from this posting is that “Seattle is not dense”, then the opponents of transit are sure to say what they always say: “Seattle is not dense, therefore Seattle can’t support Light Rail”.
This is an inaccurate statement, but this post will be used to support it.
This blog entry really should be labeled something like “Puget Sound Metropolitan Area is not dense.” That is an accurate statement supported by the data; the current title is not an accurate statement and is not supported by the data.
The average Bellevuean or Rentonian doesn’t identify themselves as being from there for people from out of state. They identify themselves as Seattleites. The only people who might object are folks from Tacoma, for historical reasons. But if they are stuck making conversation with a seatmate on the plane, do they say Lakewood, Tacoma, Seattle, or Washington state?
Incidentally, contrast that with people from Southern California (who say “Orange County”, the “Inland Empire”, or Southern California to avoid saying “Los Angeles”) or “Bay Area” people (who identify as from “The Bay” or the individual town or geographic feature, and are careful to avoid saying “San Francisco if they don’t physically live in “The City”). Seattle takes up a small percentage of Pugetopolis, but is the dominant city in the area, in contrast with the Bay Area where there are three main cities.
Ya. Lots of people want to be Seattleites, but not everyone can be. But the title of this post is still inaccurate
Take a look at the chart again. Focus on the really dense parts – 1000 sf per person or less. Although the data set is from our metropolitan area, this section is clearly only Seattle. That was one of the benefits in my mind of presenting the data this way – it separates data into the dense parts versus the sparse parts. Now, compare that line to SF or LA. We’re way under either of those. By saying “Seattle is Not Dense” I didn’t mean our region is not dense (though the data shows that too). I mean the densest parts of our region, which clearly include Seattle and Seattle only (other than perhaps a few outliers), are not dense.
Just take a step back and look at that line. We aren’t on the same playing field as SF or LA, and barely on the same planet as NYC. That’s an important point to keep in mind as neighbors start telling us we’re becoming Manhattan. We have a long, long way to go before we come close to SF or LA, let alone Manhattan.
That’s a fact. So why dilute Seattle numbers with Subuvall? If you had picked city limits (maybe more work data mining) it would have made the point even more succinct and more accurate with respect to the title of the post. I suspect it would also have shown SF as being a lot more dense than LA… which it is.
Being on planet earth is the only thing Seattle has in common with NYC from a demographics point of view. Anytime someone compares Seattle to NYC it’s pretty much game over; why even try to logically make a comparison. Really, that’s true for every other city in the USA.
Fnarf lives on Phinney Ridge. The only immigrants he sees are the Honduran nannies pushing around MSFT babies in the park.
If you folks want density so much, why not move to a dense city? I did for 15 years but moved to Seattle for a little more space.
Because some of us grew up in Seattle and have the right to say what we want our city’s future to be, and will not be driven out as if the postwar zoning mentality is somehow sacred and inviolable. Because it’s less humid than back east and the temperatures are milder. Because job opportunities are relatively good here, and the cost of living is less than in most cities with better transit. Plus the fact we’d have to leave our jobs, family, and friends.