The US Census Bureau thread

[Chicago103]

Originally Posted by Dralcoffin

Has anyone calculated what the density of Chicago would be if you took out the vast Calumet industrial zone? I'd imagine the density without South Deering (and maybe Hegewisch) would be a close model...

Also O'Hare Airport as well which is fully within city limits, Midway Airport also takes up over a square mile of land. You are right about the vast Lake Calumet area, it is quite amazing how much empty land there is down there. I have been biking down there a few times and there is a stretch of Stoney Island just east of Lake Calumet from 103rd street to the Calumet River, over three miles where you will not see a single building, its all just open fields, swamps, ponds, landfills that look like small mountains, its so surreal its hard to believe you are still within Chicago city limits. It also shows how dense the rest of the city must be to rack up those kinds of density numbers with so much empty and non-residential land.

[Ch.G, Ch.G]

Originally Posted by Chicago103

Also O'Hare Airport as well which is fully within city limits, Midway Airport also takes up over a square mile of land. You are right about the vast Lake Calumet area, it is quite amazing how much empty land there is down there. I have been biking down there a few times and there is a stretch of Stoney Island just east of Lake Calumet from 103rd street to the Calumet River, over three miles where you will not see a single building, its all just open fields, swamps, ponds, landfills that look like small mountains, its so surreal its hard to believe you are still within Chicago city limits. It also shows how dense the rest of the city must be to rack up those kinds of density numbers with so much empty and non-residential land.

Exactly. I believe that area is managed as part of the Cook County Forest Preserve system, no? They also have other sites, particularly in the northwest portion of the city. I'm not sure the total area in square miles, but it's not insignificant. When you add that to the twenty square miles of O'Hare and nearly twelve square miles of city parks (administered by the park district), you end up with a habitable area maybe 20% less than Chicago's total land area of 227 square miles.

Furthermore, if I understand Chicago's zoning, there are vast swaths of land dedicated to industrial use and are therefore uninhabitable, and vast swaths of land formerly dedicated to industrial use but which have yet to be redeveloped, sitting empty.

Calculate all this uninhabitable and uninhabited area, subtract it from Chicago's total land area, and you arrive at a population density figure much greater that the 11.8k/square mile.

Chicago is one example (and the one I know best). I'm sure, to varying degrees, you could do the same for the other cities on that list. Hence, I don't think it's safe to assume that L.A. would automatically go from tenth to second or third if one were to perform a more accurate calculation of population density on those cities.

[ChiSoxRox]

Looking at some numbers:

City of Chicago: 227.1 square miles

O'Hare community area: 32.5 sm, 11,956 pop.
Calumet area (South Deering + Hegewisch + Riverdale): 17.1 sm, 36,580 pop.
Midway airport: 1.1 sm
All parks combined: 11.9 sm

So Chicago without those areas: 164.5 sm, 2,647,062 pop.

That gives a density of 16,091 people per square mile!

[VivaLFuego]

A cleaner and less debatable way to deal with the issue of unoccupied land impacting density calcs is to just do a weighted average density calculation. Calculate the population density of each census tract, then weight that density by its total population to estimate the city total. This would tell us the density experienced by the "average" resident of that city, which it sounds like is what people are after...

EDIT: just tried doing this quickly for Chicago (including Norridge and Harwood Heights for expediency, possibly got a couple tracts along the city boundary but it should be quite close).

Weighted average population density: 19,150 ppsm
Weighted average household density: 8,858 hpsm

Cumulative Distributions

# of tracts with pop density > 60,000 ppsm: 7
# of tracts with pop density > 50,000 ppsm: 12
# of tracts with pop density > 40,000 ppsm:
# of tracts with pop density > 30,000 ppsm: 94
# of tracts with pop density > 20,000 ppsm: 286
# of tracts with pop density > 10,000 ppsm: 630
# of tracts with pop density > 5,0000 ppsm: 797
Total tracts: 927

Total population living at > 60,000ppsm: 48,610
Total population living at > 50,000ppsm: 83,301
Total population living at > 40,000ppsm: 131,829
Total population living at > 30,000ppsm: 465,150
Total population living at > 20,000ppsm: 1,227,773
Total population living at > 10,000ppsm: 2,416,829

This shows that even though Chicago's official population density is a bit under 12,000 ppsm, nearly half the city's population nonetheless lives at over 20,000 ppsm.

If you want to do this, just create a table of all tracts in your city with their square mileage and total population counts, then calculate your density and do a weighted average. Since the methodology is easily replicable, is based on hard data and city boundaries, and involves no judgment calls, it should avoid any subjective squabbling over what areas "should" be included/excluded.

[brickell]

I've been looking for someone to do the weighted density for 2010. Guess we'll have to work the numbers ourselves. I'll took at Miami.

[tech12]

VivaLFuego A cleaner and less debatable way to deal with the issue of unoccupied land impacting density calcs is to just do a weighted average density calculation. Calculate the population density of each census tract, then weight that density by its total population to estimate the city total. This would tell us the density experienced by the "average" resident of that city, which it sounds like is what people are after...

EDIT: just tried doing this quickly for Chicago (including Norridge and Harwood Heights for expediency, possibly got a couple tracts along the city boundary but it should be quite close).

Weighted average population density: 19,150 ppsm
Weighted average household density: 8,858 hpsm

Cumulative Distributions

# of tracts with pop density > 60,000 ppsm: 7
# of tracts with pop density > 50,000 ppsm: 12
# of tracts with pop density > 40,000 ppsm:
# of tracts with pop density > 30,000 ppsm: 94
# of tracts with pop density > 20,000 ppsm: 286
# of tracts with pop density > 10,000 ppsm: 630
# of tracts with pop density > 5,0000 ppsm: 797
Total tracts: 927

Total population living at > 60,000ppsm: 48,610
Total population living at > 50,000ppsm: 83,301
Total population living at > 40,000ppsm: 131,829
Total population living at > 30,000ppsm: 465,150
Total population living at > 20,000ppsm: 1,227,773
Total population living at > 10,000ppsm: 2,416,829

This shows that even though Chicago's official population density is a bit under 12,000 ppsm, nearly half the city's population nonetheless lives at over 20,000 ppsm.

If you want to do this, just create a table of all tracts in your city with their square mileage and total population counts, then calculate your density and do a weighted average. Since the methodology is easily replicable, is based on hard data and city boundaries, and involves no judgment calls, it should avoid any subjective squabbling over what areas "should" be included/excluded.

cool...but that way of doing it seems pretty time consuming. How did you do it "quickly"? Did you use the census website? I can't find any data on population density or square mileage for census tracts on the factfinder2 website...only data on population, racial demographics, and occupancy.

but here's San Francisco's density using the method Dralcoffin did:

San Francisco = 46.7 square miles (pop. 805,235/17,242.7 people per square mile)

-all parkland in SF = 8.4 square miles (source: http://www.tpl.org/content_documents...ndAgency09.pdf)
-industrial zones, port areas, the Hunter Point Naval Shipyard = roughly 3 sq. mi. (i had to eyeball this with the help of google maps). It's all lightly populated, so let's say that takes 5,000 people out of SF's total population.

So that makes a remainder of 35.3 square miles, 800,235 people, and a density of 22,669.5 people per square mile.

[fflint]

It's good to keep in mind the effect these uninhabited areas we want to exclude from our calculations have on nearby inhabited areas.

I'll use anecdote to try and get my point across, as I'm addled with Sudafed and a nasty head cold.

I once lived directly across the street from Golden Gate Park. A wall of green was all you saw from my third-floor bay window. My census tract was somewhere in the 30k ppsm range, but there was zero density across the street--and that had a big effect on the area. From that uninhabited place came zero additional through-traffic, zero additional competitors for a parking spot or a spot in line at the grocery store, zero additional customers for goods and services, etc. Indeed, the grocery store was the only notable commercial establishment. My area was walled in by a green curtain through which the masses did not pass.

I now live in a more central neighborhood, through which passes nearly every Muni Metro line as well as two BART lines and the city's main drag. My tract is roughly 28k ppsm (2000 numbers, haven't seen the new ones yet)--slightly less than the prior area--but is surrounded on all sides, and for long distances, by tracts of equal or higher population density. There is tremendous through-traffic on foot, bike, rail, bus and by car. All that through-traffic leads to some notable additional demand for goods and services, additional demand for parking, additional competition for a place in line at the grocery store. It is a more commercial area than my prior neighborhood because that increased demand can sustain more activity. There are more people around at all times. There are 24-hour establishments. Yet if you go only on the density numbers, my old place fronting the green curtain was "more dense."

I experience the effects of a much higher density where I live now, however, precisely because it does not abut an uninhabited zone. We can exclude empty land, greenbelts, parks, and industrial zones from our density calculations to get at the density "the average resident experiences"--but that isn't going to tell the whole story. Vast, uninterrupted stretches of high density are, IMO, going to seem more dense to residents than chunks of equally high density abutting uninhabited areas.

[tdawg]

Wow, I live in Queens and simply cannot accept that the county grew by only 1,600 people over the past decade. That many people have moved to my neighborhood (Astoria) in less time than that.

[dktshb]

Originally Posted by Ch.G, Ch.G

Exactly. I believe that area is managed as part of the Cook County Forest Preserve system, no? They also have other sites, particularly in the northwest portion of the city. I'm not sure the total area in square miles, but it's not insignificant. When you add that to the twenty square miles of O'Hare and nearly twelve square miles of city parks (administered by the park district), you end up with a habitable area maybe 20% less than Chicago's total land area of 227 square miles.

Furthermore, if I understand Chicago's zoning, there are vast swaths of land dedicated to industrial use and are therefore uninhabitable, and vast swaths of land formerly dedicated to industrial use but which have yet to be redeveloped, sitting empty.

Calculate all this uninhabitable and uninhabited area, subtract it from Chicago's total land area, and you arrive at a population density figure much greater that the 11.8k/square mile.

Chicago is one example (and the one I know best). I'm sure, to varying degrees, you could do the same for the other cities on that list. Hence, I don't think it's safe to assume that L.A. would automatically go from tenth to second or third if one were to perform a more accurate calculation of population density on those cities.

Here is 49.7 square miles of LA from Downtown outward (greater than the city boundaries of Boston, Miami and San Francisco). According to city data zip code 2009 estimates LA comes in at 20,261 people per square mile :

Zip sq mi density Population
90011 4.4 23859 103882
90014 0.3 12478 3610
90015 1.7 9397 15533
90017 0.7 28577 21234
90007 2.8 16665 46207
90057 0.9 51143 45145
90005 1.1 38507 44148
90004 3.1 22781 69638
90026 4.2 18219 75613
90029 1.4 42796 42796
90028 1.5 21072 31367
90038 1.5 21648 33415
90018 2.9 16660 48369
90016 3.6 13391 48206
90006 2 33012 64419
90019 3.9 17539 69289
90036 2.5 xxxxxx 33752
90020 1.1 xxxxxx 43500
90048 1.9 xxxxxx 22296
90037 2.8 xxxxxx 58185
90062 1.9 xxxxxx 30050
90001 3.5 xxxxxx 56362
49.7 1007016
20,261 people per square mile.

http://www.city-data.com/zipmaps/Los...nia.html#90001

Granted this is not 2010 census numbers but you get an idea of how LA's core is denser than places like Miami and Boston. It most certainly is denser than Washington DC and Long Beach and most likely Philadelphia. I am not sure about San Francisco though even though the numbers indicate it I know this is 2009 estimates and not 2010 census numbers but I still don't think the average would drop by over 3092 people per square mile.

[blade_bltz]

^ but why are you assuming that all cities must have 50 sq mi "cores". Boston's "core" is certainly not identical with its ~50 sq mi (48.4) city limits, which contain several highly suburban neighborhoods. Since the Boston metro is dwarfed by LA's, why would you expect their cores to be same size?

[dktshb]

Originally Posted by blade_bltz

^ but why are you assuming that all cities must have 50 sq mi "cores". Boston's "core" is certainly not identical with its ~50 sq mi (48.4) city limits, which contain several highly suburban neighborhoods. Since the Boston metro is dwarfed by LA's, why would you expect their cores to be same size?

Sure, I suppose it goes both ways I just stopped at 47 square miles because I thought there was no point to continuing.

[Shawn]

^ Right on. You'd include Hyde Park and West Roxbury, both inner-ring suburbs which happen to fall inside Boston's municipal borders, while excluding Cambridge, Somerville, Chelsea and Everette right across the Charles and Mystic. I'm pretty certain all of Cambridge is physically closer to the Statehouse or whatever you want to use as a "core" marker than anywhere in Westie or Hyde Park.

[lawfin]

Originally Posted by VivaLFuego

A cleaner and less debatable way to deal with the issue of unoccupied land impacting density calcs is to just do a weighted average density calculation. Calculate the population density of each census tract, then weight that density by its total population to estimate the city total. This would tell us the density experienced by the "average" resident of that city, which it sounds like is what people are after...

EDIT: just tried doing this quickly for Chicago (including Norridge and Harwood Heights for expediency, possibly got a couple tracts along the city boundary but it should be quite close).

Weighted average population density: 19,150 ppsm
Weighted average household density: 8,858 hpsm

Cumulative Distributions

# of tracts with pop density > 60,000 ppsm: 7
# of tracts with pop density > 50,000 ppsm: 12
# of tracts with pop density > 40,000 ppsm:
# of tracts with pop density > 30,000 ppsm: 94
# of tracts with pop density > 20,000 ppsm: 286
# of tracts with pop density > 10,000 ppsm: 630
# of tracts with pop density > 5,0000 ppsm: 797
Total tracts: 927

Total population living at > 60,000ppsm: 48,610
Total population living at > 50,000ppsm: 83,301
Total population living at > 40,000ppsm: 131,829
Total population living at > 30,000ppsm: 465,150
Total population living at > 20,000ppsm: 1,227,773
Total population living at > 10,000ppsm: 2,416,829

This shows that even though Chicago's official population density is a bit under 12,000 ppsm, nearly half the city's population nonetheless lives at over 20,000 ppsm.

If you want to do this, just create a table of all tracts in your city with their square mileage and total population counts, then calculate your density and do a weighted average. Since the methodology is easily replicable, is based on hard data and city boundaries, and involves no judgment calls, it should avoid any subjective squabbling over what areas "should" be included/excluded.

Don't suppose you could do this for north side only; and northside lakefront from Near North to RP?

[fflint]

Originally Posted by dktshb

LA's core is denser than places like Miami and Boston.

First of all, this is a comparison of apples to oranges.

Second, you're defining "LA's core" subjectively as "All the most dense areas I can find that can be put together in an arbitrarily-sized land area." Is that really what a core is?

Third, is cherry-picking the most dense 50 square miles for comparison with Boston any more legit than cherry-picking the least dense 50 square miles in LA for comparison with Boston? Because this game can be played both ways.

[dktshb]

No cherry picking... I went from all of downtown outward in a continuous pattern and not jumping looking for denser areas. Yes I could have made it less dense focusing on industrial areas to the east and bypassing certain tracts to get to the santa monica mountains but that wouldn't be continuous and I don't think anybody would dispute that the area I chose is not the core of LA which streches from Downtown to Hollywood out to beyond Fairfax and south of the 10, which is the area I covered. The city data also didn't do the san fernando valley or the area around the airport so I couldn't do them if I wanted to. The point was to show LA has comparable densities to these cities. I do understand this still isn't an exact science but the point was LA is just as dense as the most dense large cities in America. You know very well there is an allowed misrepresentation about how dense the city really is and that really needs to stop. I am all for certain guidelines to really have a better less subjective and more objective way to calculate a city density. I think you would agree holding LA's density to its square mileage to these cities 2 to 6 times smaller in land area is not realistic and is also comparing apples to oranges too. I know that Boston's whole city limits includes parks, airport, etc. which isn't included here and I am all for a set way to do this.

[Chase Unperson]

Originally Posted by fflint

First of all, this is a comparison of apples to oranges.

Second, you're defining "LA's core" subjectively as "All the most dense areas I can find that can be put together in an arbitrarily-sized land area." Is that really what a core is?

Third, is cherry-picking the most dense 50 square miles for comparison with Boston any more legit than cherry-picking the least dense 50 square miles in LA for comparison with Boston? Because this game can be played both ways.

If the area is continuous, it is a meaningful statement. It would imply that LA could contain all of Boston, land and people, as an intact proper subset. It makes a meaningful comparison between LA and Boston. LA can not contain NYC as a proper intact subset. That is a meaningful comparison between NYC and LA.

If you take the least dense 50 square miles and compare it to Boston. It would mean Boston could contain a subset of LA. That doesn't really mean anything at all in terms of comparing LA and Boston.

[Quixote]

Sorry for taking the easy way out...

Originally Posted by dweebo2220

I've always felt like there was a distinct feel to the area of LA that was once served by streetcars (I'm not talking about the interurban red cars, but the actual "zoned" network seen below in a map from the 40s).


I recently wondered what this area's stats would be if taken as it's own city, since it seemed to me to have a lot in common with older "traditional" American cities. I found a website today that lists demographic info (including area and population) for all of the census-designated areas in the city of LA, and I found it was really easy to add up census tracts and, in effect, draw my own city borders.

The map below shows this area as its own city.. I'm calling it "Classic" LA but I'm open to other names.



It's interesting that the census tract ranges, shown on the map with different colors, are pretty much contiguous with the different greater neighborhoods like koreatown, boyle heights/lincoln heights, downtown, etc.

I know L.A. as a whole is not a "traditional" city, but this 128 sq. mile area pretty much is. It's almost exactly the same land area as the city of Philadelphia yet has almost 300 thousand more people. Of major US cities, it would have the third highest population density after New York and San Francisco (I know, I know, this is mostly due to overcrowding..).

As a percentage of the entire LA-Long Beach-Santa Ana urbanized area, it's 15% of the total population and 7.5% of the total land area. Interestingly, the cities of Boston and Washington D.C. also contain about 15% of their respective area populations.

It would be great if this thread could start a discussion about this unique area of the city. I don't want to pretend like LA is "traditional"--it's growth and development is unique among US cities--but I would like to point out the fact that there is still a rather traditional city right smack in the middle of LA, if you know how to look at it.

[The ATX]

There is some interesting information regarding Austin's land area and population on the City of Austin's Demographics and Annexation Websites.

Austin annexed a fair amount of populated areas after the 2010 census count - mostly in December 2010. The city estimates the 04/01/2011 population at just over 812,000. That's 22,000 more than the official census count. In addtion to that, there are more than 208,000 people living in Austin's ETJ (Extraterritorial Jurisdiction). If Austin went on an annexation binge and didn't allow the ETJ subdivisions to incorporate, the city's population would be 1,020,000 right now without any more growth.

I know many boosters of Northern cities don't like the fact that Sunbelt cities can annex population. I prefer denser central cities as well. But suburban growth is going to happen no matter what is happening in the central city. The fact that a city can annex and add to its tax base instead of a bunch of little suburbs taking control of suburban growth is a very good thing.

http://www.ci.austin.tx.us/demographics/

http://www.ci.austin.tx.us/annexation/compannex.htm

[fflint]

Originally Posted by dktshb

No cherry picking... I went from all of downtown outward in a continuous pattern and not jumping looking for denser areas.

Cherry-picking data:
75% of the land area of downtown LA is within area codes 90012, 90013, and 90021 with a sliver of downtown in 9033 (which contains Boyle Heights). Yet all four of these downtown zip codes are inexplicably excluded from your exercise, and the likely reason is they are mostly non-residential. That is cherry-picking data.

Subjective definition:
How can one legitimately define LA's "core" as containing only 25% of downtown and excluding Boyle Heights while including the Fairfax?

[Chase Unperson]

Originally Posted by fflint

Cherry-picking data:
75% of the land area of downtown LA is within area codes 90012, 90013, and 90021 with a sliver of downtown in 9033 (which contains Boyle Heights). Yet all four of these downtown zip codes are inexplicably excluded from your exercise, and the likely reason is they are mostly non-residential. That is cherry-picking data.

Subjective definition:
How can one legitimately define LA's "core" as containing only 25% of downtown and excluding Boyle Heights while including the Fairfax?

That is a good point.

However even if downtown is excluded a 50 square mile contiguous area that contained a million people would still imply that Los Angeles could contain an intact Boston or San Francisco, Land and People as a proper subset.