The US Census Bureau thread

[ChiSoxRox]

Originally Posted by KevinFromTexas

So, does anyone have the link for the city, county, and metro 2020 numbers? I'll add them to the first post of the thread for easy access...

The raw population numbers? I've updated the Wikiprdia lists on cities and metro areas, although not yet the county list.

[wwmiv]

Originally Posted by ChiSoxRox

Weighted population density is defined as the sum of the products of population (n) and density (ρ) for a tract, normalized to the entire population (N),

i.e. WPD = Σ (n*ρ) / N

Put another way, it is summing up the population of each tract (n) weighted by that tract's share (f) of the total population

WPD = Σ (n*f)

Going a little bit deeper:

This method is arithmetic mean (or average) weighted density and the proper interpretation of the output is:

The average person in the City of Chicago lives in a census tract with 21,235 people per square mile.

If you used another method, such as median weighted density, the interpretation would be slightly different:

The median person in the City of Chicago lives in a census tract with XYZ people per square mile.

Or geometric mean weighted density, where the underlying distribution is skewed, which has the same interpretation as arithmetic mean weighted density.

https://www.worldpop.org/methods/pwd

[KevinFromTexas]

Originally Posted by ChiSoxRox

The raw population numbers? I've updated the Wikiprdia lists on cities and metro areas, although not yet the county list.

That's cool, thanks, but I was also wanting all of the podunk places, too.

[ChiSoxRox]

Originally Posted by KevinFromTexas

That's cool, thanks, but I was also wanting all of the podunk places, too.

Citypopulation.de has done a very extensive job of scraping the data. They also have really nice density, change, etc. maps.

[benp]

This is a good site to find data for all locations, and includes 2010-2020 comparisons.

https://data.dispatch.com/census/

[Yuri]

Originally Posted by KevinFromTexas

So, does anyone have the link for the city, county, and metro 2020 numbers? I'll add them to the first post of the thread for easy access...

Go to City Population: http://citypopulation.de/ . The US section: http://citypopulation.de/en/usa/

I find extremely friendly to navigate there. You can go down census tracts, with numbers going back to 1990 Census.

[benp]

Originally Posted by ChiSoxRox

Weighted population density is defined as the sum of the products of population (n) and density (ρ) for a tract, normalized to the entire population (N),

i.e. WPD = Σ (n*ρ) / N

Put another way, it is summing up the population of each tract (n) weighted by that tract's share (f) of the total population

WPD = Σ (n*f)

With the census tract numbering from Seattle in mind, I checked Los Angeles. 1031, 1732, 2309, ah the pattern doesn't hold. Wait, Beverly Hills is in the 7000s, Long Beach tracts are 57xx, and so on. What happens if I sum up the 1xxx and 2xxx tracts for Los Angeles County? Those seem to map to LA city limits! The total population in those tracts is 3,898,921. The city of Los Angeles is 3,898,747. Those tracts combined are less than 200 people away! A typical tract is 3k-4k people or so, so that cannot be a coincidence. I've checked the edges of LA city limits, and the 1xxx/2xxx tracts map to the city of Los Angeles! I've cracked the numbering code!

Edit - Turns out that depends on the city. For example, San Diego does not follow the self-partitioning pattern.

Weighted densities for cities proper
New York, NY…...65,299.0 ppsm
Jersey City, NJ…...36,846.2
San Francisco, CA…...33,572.5
Boston, MA…...27,437.4
Philadelphia, PA…...21,935.1
Chicago, IL.......21,235.2
Washington, DC…...20,642.9
Arlington, VA…...19,960.8
Los Angeles, CA…...17,294.3
Alexandria, VA…...15,708.3
Seattle, WA…...15,249.3
Baltimore, MD…...11,333.0
Milwaukee, WI…...9,506.5
Denver, CO…...8,889.4
New Orleans, LA…...7,189.8
St. Louis, MO…...6,834.8
Richmond, VA…...6,413.2

Thanks - Using that definition I ran the numbers for my city and its densest suburb to see how it aligns with the above.
Buffalo, NY ......9,265.5
Kenmore, NY ......10,718.8

[Steely Dan]

Milwaukee, WI…... 9,506.5
Buffalo, NY...... 9,265.5

More evidence for the whole "cousins at opposite ends of the lakes" theme.

[Handro]

Originally Posted by Steely Dan

Yep.

It's why I have long maintained that WPD for Urban Areas is likely the closest we'll ever get to an apples-to-apples density comparison for US cities. None of the arbitrariness of city limits, and none of the absurdity of the MSA county mash-up game.

UA's ain't perfect either, but we ain't never gonna get perfect.

Is there anywhere to see the weighted densities for urban areas from the 2010 census?

[dimondpark]

Yes LA, come thru

[DCReid]

Are these the correct 2020 population numbers for metros?

https://en.wikipedia.org/wiki/List_o...tistical_areas

[Steely Dan]

Originally Posted by DCReid

Are these the correct 2020 population numbers for metros?

https://en.wikipedia.org/wiki/List_o...tistical_areas

considering that SSP's own master population statistician, ChiSoxRox, uploaded the new MSA figures to wikipedia himself, yes, those are the correct MSA population for Census 2020.

[subterranean]

Originally Posted by DCReid

Are these the correct 2020 population numbers for metros?

https://en.wikipedia.org/wiki/List_o...tistical_areas

I've noticed at least one figure that I'm questioning (probably others). The Lansing-East Lansing MSA in 2010 was 464,036, but the 2020 MSA page linked here says that the 2020 and 2010 populations as 541,297 and 534,684, respectively.

[ChiSoxRox]

Originally Posted by subterranean

I've noticed at least one figure that I'm questioning (probably others). The Lansing-East Lansing MSA in 2010 was 464,036, but the 2020 MSA page linked here says that the 2020 and 2010 populations as 541,297 and 534,684, respectively.

Are you using the most recent definitions? The numbers on that page are using the January 2020 MSA definitions. The extra 70k is probably Shiawassee County.

Originally Posted by benp

Thanks - Using that definition I ran the numbers for my city and its densest suburb to see how it aligns with the above.
Buffalo, NY ......9,265.5
Kenmore, NY ......10,718.8

Thanks, I'll add it to the list. I'm thinking of having the city proper list cut off at 200,000 people and 6,000 ppsm weighted density. That way it should be completable this weekend.

Originally Posted by Steely Dan

considering that SSP's own master population statistician, ChiSoxRox, uploaded the new MSA figures to wikipedia himself, yes, those are the correct MSA population for Census 2020.

Aw, thanks! Although that reminds me, I (or another user) have to stitch the CSA populations together to update the CSA list still.

[iheartthed]

I tried to calculate the weighted census tract density for Detroit, Highland Park, and Hamtramck. I pulled the numbers for Detroit manually, so there might be a fat finger error or two. HP and Hamtramck were easier since there were fewer tracts:

Detroit 6,250 ppsm
Highland Park 3,868 ppsm
Hamtramck 16,904 ppsm

[ChiSoxRox]

Originally Posted by iheartthed

I tried to calculate the weighted census tract density for Detroit, Highland Park, and Hamtramck. I pulled the numbers for Detroit manually, so there might be a fat finger error or two. HP and Hamtramck were easier since there were fewer tracts:

Detroit 6,250 ppsm
Highland Park 3,868 ppsm
Hamtramck 16,904 ppsm

Thanks! I was curious to see where Detroit would land with respect to St. Louis, so about 90% of St. Louis' WPD.

Looks like mapping city limits to census tracts is a highly variable process. Los Angeles was one of the more pleasant cases. Miami is another very dense city proper, but the tract numbering in Miami-Dade goes continuously through the county, meaning I'll have to assemble Miami city tract by tract tomorrow (unless someone beats me to it). Edit: In fact, I probably won't get an exhaustive city list done. Too tricky to pull out tracts (if not impossible, looking at you Houston), and city limits are too apples to oranges for my liking.

But as a condensed version, here's the 200k club for BosWash, with outliers both above and below.

New York, NY......65,299.0
Jersey City, NJ......36,846.2
Boston, MA......27,437.4
Newark, NJ......24,478.2
Philadelphia, PA......21,935.1
Yonkers, NY......21,574.1
Washington, DC......20,642.9
Arlington, VA......19,960.8 (technically a county, I know)
Baltimore, MD......11,333.0

Baltimore's city limits do include a sizable ring of more suburban environments, but it's also the only member of this list still dropping. Perhaps the core of Baltimore is just that hollowed out.

Edit: Turns out Worcester, MA is back above 200k with a WPD of 9,549.4 ppsm, although most of Worcester is much more suburban in character.

[subterranean]

Originally Posted by ChiSoxRox

Are you using the most recent definitions? The numbers on that page are using the January 2020 MSA definitions. The extra 70k is probably Shiawassee County.

Ahh, I didn’t think they retroactively amended their definition for the 2010 census. It makes sense for continuity purposes but I never thought they’d be that sophisticated, lol.

[Steely Dan]

Originally Posted by iheartthed

Highland Park 3,868 ppsm
Hamtramck 16,904 ppsm

these two make for such an interesting case study in how two similarly-sized burbs enclaved within a much larger city, and located right next door to each other, can have such radically divergent fates over 3 decades.


population in 1990:

Highland Park: 20,121
Hamtramck: 18,372


population in 2020:

Highland Park: 8,977 (-55.4%)
Hamtramck: 28,433 (+54.8%)


i mean, talk about two completely opposite trajectories.

what did hamtramck do right?

what did highland park do wrong?

[iheartthed]

Originally Posted by Steely Dan

these two make for such an interesting case study in how two similarly-sized burbs enclaved within a much larger city, and located right next door to each other, can have such radically divergent fates over 3 decades.


population in 1990:

Highland Park: 20,121
Hamtramck: 18,372


population in 2020:

Highland Park: 8,977 (-55.4%)
Hamtramck: 28,433 (+54.8%)


i mean, talk about two completely opposite trajectories.

what did hamtramck do right?

what did highland park do wrong?

The short answer is immigration. Hamtramck has pretty much always been an immigrant magnet. Highland Park was far more directly attached to the auto industry, first with Ford and later with Chrysler. Both eventually abandoned the town they either created (Ford) or kept alive (Chrysler) and there was nothing to fall back on.

But even knowing that it's still shocking how much they've diverged. In 1990, I think most people would've considered Highland Park to be in far better shape than Hamtramck.

[ChiSoxRox]

Looking for a peer to NYC, I've used the arrondisements of Paris as the subdivisions for the city's weighted density (lopping off the Bois de Boulogne and Bois de Vincennes).

New York City…...65,299.0 ppsm

Manhattan…...108,042.0 ppsm
Bronx…...72,874.9
Brooklyn…...60,275.6
Queens,…...46,619.3
Staten Island…...15,079.8

Paris proper…...69,812.2 ppsm

The different scales (20 divisions for Paris, 361 for the Bronx over similar areas) means the Paris number is somewhat lower than a fair comparison would be, but still I find it a worthy comparison.

Edit: Found the Paris census tract equivalents. Maybe this weekend...

I'll crunch the city of Miami's weighted density tomorrow, and then probably call it good on the census calculations for now. In hindsight, IMO the most telling figure is the MSA weighted densities, especially since weighted density mostly negates the empty collar county problem.