Rich and Strange Aeons

Over the years I've done various density and walkability calculations, estimating that the required density could go as low as 4000 people/km2 (8000 people in a 12 minute walk) or even 3000 people/km2 (9000 people in a 15 minute walk.) But in the literature, messy as it is, it seems more like 10 du (dwelling units)/acre is the expected minimum, which at 2.6 people/du is more like 6500 people/km2. So I want to poke at my assumptions.

Summary: yeah, my old assumptions were flawed, and I'm now looking at closer to 10,000 people/km2 for good walkable density. Data indicates you start getting more walking before that, like 5000/km2, but it levels off above 10,000, possibly because all the trips it is easy to make walkable, have been made walkble. And per older posts, you can reach these densities with single-family housing if you insist, though you'll need to accept small lots and yards.

( Read more... )

There are other benefits to higher density, of course: more taxpayers to pay for infrastructure, more riders to justify high transit frequency, letting more people live close to attractive points like subway stations, letting people have more and more interesting lives in walking or biking distance. But in terms of reducing car trips in favor of walking, the low-hanging fruit gets plucked by 10,000.

Revisiting yesterday's post in American units, since I mostly want to persuade Americans.

Thesis: low-density living, with single family houses and sizable yards, is compatible with low-car if not car-free living, if you go in heavily on bicycles.

( Read more... )

Various sources point to a minimum level of population density needed for walkability. A source I have lost said 10-20 dwelling units (du) per acre. This Australian model derived 25 du/hectare (2500 du/km2), which is the same as 10 du/acre, as a minimum, though 35 was notably better. At an assumption (as the paper used) of 2.6 people per du, 25 du/hectare is 6500 people/km2, 35 is 9100. My own personal experience, of places I have lived and looked up the densities of, is that nice walkability starts around 9000 people/m2, while 6000 tends to be doable but a bit anemic.

( Read more... )

And in reality, there is no reason for housing to be so uniform. Left alone, people would naturally build taller and live denser near high value locations like train stations, so walkability can be supported by a mix of SFH and multifamily/rental housing. But it's good to know that you can support it with pure SFH too... as long as you allow small lots.

Though it also means that bigger lots that don't support bigger households (via large household or various rental units) are kind of free-riding on higher density elsewhere, if the inhabitants enjoy walkability.

What is the most efficient transportation use of street space, e.g. in moving or serving people per hour? I've recently seen various infographics that had numbers, but no sources or calculations. Much better is this Urbanist article, which actually gives its assumptions and calculations. But I feel like doing my own estimates, though I will draw on that article, plus other research I've done. I'll mostly be analyzing a single 3-3.5 meter lane, in city conditions (lots of intersections, thus signalled to flow only half the time.)

( Read more... )

So, there you have it. Granting that in modern society, we'll have at least one car lane on our streets, it's still a valid question as to whether an additional lane should be used for car movement or something else. And in fact almost any other use, except car parking, can serve more people. Probably the easiest approach is to split a lane's worth of space between a bike path and a wider sidewalk, but a bus lane or light rail track would be good as well. If you have two lanes' worth of space, at least at strategic places for BRT stops, then you can have really high BRT capacity.

I have a longer thing in my head but I am Busy so want to get the core ideas out so they stop bouncing around.

There are more complex approaches, but a simple one is that you're safe if there's no one infected in the room with you.

Say you dine out in a small place, exposure to 10 people. Also that you want to be safe as a medium-term habit, say over 100 meals, so 1000 people. "Safe" meaning 90% chance of no exposure to infected people, so you want community infectiousness to be 1 in 10,000 people. Nobody measures that, so assume an infected person is infectious for 10 days on average, so you want daily new infections to be 1 in 100,000, or 10 per million.

A few countries kept such levels before omicron or at least delta. Now, though, I infer case rates on the order of 1000 per million per day. Or maybe 10,000.

Some dining areas like cafeterias or dim sum palaces can seat 200 people at at time; poor ventilation could mean that by dinner time 1000 people had been breathing the air before you; eating out for life could mean 10,000 meals in your future.

Bon appetit!

I've seen various people say "everyone's going to get omicron" -- usually in a smug way, along with "covid is endemic", implying that it's foolish to try to avoid it. Usually these people are healthy and not obviously possessed of frail people they care about. At any rate, I'm inclined to try to prove them wrong, and avoid getting it until they roll out new vaccines or prove that it's really "just a cold" or something. Which means, basically, no unmasked contact with people until rates have come down a lot. But come down how much?

( Rambling )

I'd wondered why child care is so expensive yet pays the workers so little. I spent part of yesterday reading articles and watching videos; the answer seems pretty simple.

Average US fees are like $10,000 per kid. So say a worker is caring for 3 kids. That's $30,000, to pay for everything. Even if overhead is very low, so that 80% goes to compensation, that's just $24,000 to pay salary and payroll tax and (ha ha) benefits. If labor gets just 60% (still high compared to most other retail business), we're down to $18,000.

If there are 5 kids, those numbers become $50,000 gross, $40,000 at 80% labor, $30,000 at 60%. But quality of care is thought to drop, and many states require 3-4 kids per carer.

Mississippi allows 5 to 1, though it seems workers get paid the same low rate; instead parents only have to pay $5000 for child care. (Source) Of course rents matter for overhead -- Massachusetts requires 3 to 1, and the average fee is more like $16,000. (Though some quick searches claim pay is more like $30-35k, of course that has to pay MA, largely Boston, rents as well.)

Do the numbers check out? MS: 5x$5000 = $25,000. Searches claim $17,000, $22,000, $25,000 -- latter may be *teacher*, vs. an aide; one video suggested 40 kids might have 3 teachers and 6 aides. MA: 3*$15,000 = $48,000, allowing $28-38,000 in labor.

Going the other way: if a worker makes $30,000/year ($15/hour minimum wage, we'll ignore what daycare hours are actually like), and has 4 kids, and overhead is 25%, fees have to be 30/0.75/4 = $10,000/year. Except I forgot payroll taxes, so probably more like $11,000.

So yeah. it's pretty intrinsically expensive. How do other countries offer free child care? It's possible they skimp on the ratios, but mostly it must be because the government spreads the cost over all taxpayers. Instead of a family paying $10,000 for a few years, everyone's paying like $1000 (*very* averaged) every year.

(If a kid needs care for 5 years, $50,000 total, and pays taxes for 40 years, as an adult they pay $1250/year of paying taxes.)

I still like my 2019 model of how even 3000 people/km2 can be kind of walkable (and bikable) if you do it right. Short recap: supermarket needs 8-12,000 people, so imagine a 2 km x 2 km superblock, with a market right in the center (along with bus/rail intersection). Furthest walk is 2 km or 24 minutes, from the corners, which isn't great, but many people have shorter walks, especially if density is non-uniform and clumps by the center. And 2km is 10 minutes on a slow bike, so *that's* nice. I also penciled in schools and stores and such.

At 12,000 people/km2 the scaling is easy: you have the same pattern, but in a 1x1 km superblock. Now the longest walk is 12 minutes, super easy, barely an inconvenience.

But what about 6,000? I found that annoying to think about and didn't talk about it last time. Annoyingly supermarkets might need just 8000 people... but that's 1/3 more than I have in 1 km2.

But I realized, if you take the 12,000/km2 model, and then simply remove every other supermarket in a checkerboard pattern, then you have the right population ratio to support the markets, but still no one has to walk more than 12 minutes, 1 km. You've increased the *average* walk time -- a bunch of people who were in 0-3 minutes are now in 9-12 minutes -- but not the max.

Probably not coincidentally, I read somewhere that car use drops off around 10-20 units per acre; if that's gross acre, then that's 2500-5000 units per km2, or 6250-12,500 people/km2 if each unit averages a 2.5 person household.

(And if you're wondering how that compares to real cities, this older post is useful, though sometimes neighbhorhood density would be more useful. But given that cities tend to be largely residential, a US city of 1000-2000 people/km2 is obviously not going to have much walkability.)

Something I've been struggling with recently: my various models of a city built around cars yield numbers similar to actual densities. My models of human oriented cities yield much higher densities than the real cities that inspire them. But I think I've realized at least part of why.

To recap:

Car city model: assume 1/3 is streets/roads, 1/3 is zoned residential, 1/3 commercial. Assume half the commercial land is surface parking, at 30 m2/space. A model km2 thus can have 1e6 m2/6/30 m2/space = 5555 parking spaces; at 3 non-residential spaces per car, that's 1851 cars. Multiply by 1.25 for non-drivers, and get a population density of 2300 people/km2, which is about as dense as post-war cities get in the US.

People city model: 1/3 street, and the rest with an average residential FAR of 2.0, whether that's residential neighborhoods with houses or mixed/commercial neighbhoods with housing above shops. That's 1.3e6 m2 of living space per km2, allowing 33,000 people at 40 m2 per person (reasonable to me) or 16,600 people at 80 m2 per person (current US average).

Osaka and Tokyo, which at least *look* like they should be hitting FAR 2 or greater -- lots of 2 story houses filling their lots, lots of high rises -- are 12,000 and 15,000 people/km2, at only 19 m2 per person. That's off by a factor of 4.

But my people model assumes no cars at all, or that they can treated as trivial. In fact Japan supposedly has a lot of cars, 0.6 per person nationwide. Assuming only 2 parking spaces per car, that's 0.6*2*30 = 36 m2 of parking per person, while nationwide there are 22 m2 of living space per person. So there's as much, or more, parking than living space, even in Japan. The cities would have fewer cars but also less space per person.

And that's cars parked in multi-level garages taking half of the built up floor area. Much of the urban parking is actually open surface lots, whether small commercial lots in neighborhoods or parking attached to stores. When all the surrounding building is lot-filling 2+ story buildings, each 30 m2 of surface parking displaces 60 or more m2 of floor space.

So there's a factor of 2. It's also possible I overestimate the average residential FAR, not accounting for industrial zones or parks and shrines or overestimating mixed buildings, I dunno.

This adjusted model implies that a huge chunk of Tokyo buildings are parking, which wasn't my impression, but also wasn't something I was paying attention to. Of course, I was also mostly in places not far from train lines.

Hmm, this says that in new Tokyo (23 wards) condominiums, the ratio of parking to spaces is 30% -- 2064 spaces for 7008 apartments. But in 2007 the ratio was 56%. I don't know how big these condos are: studios, 2BR, what? Or how many people are in each one. But an average of 15 m2 of parking attached to each not very big condo is a fairly sized chunk.

Caveat: Japanese cars are smaller; there's also robotic parking that I assume takes less space overall. Houses can have parking lots that go directly to the street and thus don't need access lanes, though these are often open surface parking that displace multi-story density above them.

This is tangentially fascinating -- cities limited wheeled carts even before cars, with most transport by canal; most canals were later filled in to make arterial roads; less than 2% of Japanese streets are wider than 5.5 m, and 35% are too narrow for even one car.

I have failed to find how many parking spaces Japan or its cities have.

This comes up a lot in discussion of density. The idea being that 2-4 story or such buildings are "human scale", vs. skyscrapers. My gut has some sympathy with it. OTOH...

A few years ago I took a bus to NYC. As we crawled down the avenues of Manhattan, I looked out at the 18 story or whatever buildings and felt intimidated.

But once I was living there, walking around those buildings was usually fun. What you see at ground level is a lot more salient than what's above, and lots of narrow businesses provide tons of options and stimulation; if it's more than human scale, it's less from height and more from the sheer overabundance of options. Also a poor people/sidewalk ratio.

Likewise, when I first visited central Tokyo, after some years in Indiana, in a way I felt at home, despite the different culture and language. Tall buildings, busy-ness of shops and pedestrians, knowing there was a good transit nearby... it felt more human *friendly*, anyway, than even downtown Bloomington.

Tangent on how many options: say you're standing on a Manhattan avenue, and are willing to walk 10 minutes, or 1000 meter. If stores are 4 meters wide, that's 250 stores. x2, because often two stories of businesses. x2, because both sides of the street. And x2 again, because you can walk 10 minutes north or south. So 2000 small businesses in a 10 minute walk! Well, no businesses in intersections, so take off 1/4 or even 1/3. And some are probably bigger, 6 m or more. 2000 * 2/3 * 4/6 = 888 businesses... still a lot. And not counting anything that might be on cross-streets, or walking over to an adjacent avenue.