The US Census Bureau thread

[Yuri]

Originally Posted by ChiSoxRox

Tracts are divided into block groups, which in turn are divided into blocks.

However, for urban areas, sub-tract divisions get *very* numerous very quickly, so tracts strike a decent balance between granularity and amount of data.

You're either very brave or have better techniques than I do. If I want to look into a region with over 30 census tracts, my anxiety kicks in. And I use City Population with their very friendly maps. However, I have to copy and paste one by one to add up on the Excel. I don't know to do it differently.

[ChiSoxRox]

Originally Posted by Yuri

You're either very brave or have better techniques than I do. If I want to look into a region with over 30 census tracts, my anxiety kicks in. And I use City Population with their very friendly maps. However, I have to copy and paste one by one to add up on the Excel. I don't know to do it differently.

I use CSVs, popping into Excel (with perhaps a bit of formatting clean up in a text editor).

I'm digging around citypopulation seeing if they have an "export CSV" button. I'd love to get the weighted density of Paris using their equivalent of tracts but that's not happening if I have to go by hand.

[ChiSoxRox]

May 26 is when we get 2021 city population estimates. Same time frame as the county estimates from March, so they're going to be ugly from the peak of WFH.

[Yuri]

Not sure if someone mentioned previously, but the US mean center of population barely moved. In fact, only 12 miles, the shortest distance since the first Census back in 1790:


Wikipedia

Those are the growth rates for the US and the four major regions between 1990 and 2020:

UNITED STATES ---- 7,35% ---- 9,71% --- 13,15%

Northeast ----------- 4,14% ---- 3,21% ---- 5,48%
Midwest ------------- 3,08% ---- 3,94% ---- 7,92%
South -------------- 10,22% --- 14,29% --- 17,31%
West ----------------- 9,23% --- 13,84% --- 19,72%

The West has been growing slower than South since the 2000's, for the first time in history. The Northeast went against the trend and grew faster in the 2010's than the 2000's. All that is contributing to the slowdown.

And as the West has never been so close to the national average, it's not inconceivable that for the first time ever, the mean centre won't move any inch westwards, only southwards, specially now as it's already very close to Texan eastern border.

Ironically, it seems the "western" California that was the one pushing the centre southwards back for good part of the 20th century. Now it's the "southern" Texas that is the main force making the centre still keep moving westwards.

[JoninATX]

The U.S. census estimates it missed more than a half-million Texans during 2020 count

Tripped up by politics and the pandemic — and with only a last-minute investment in promotion by the state — the 2020 census likely undercounted the Texas population by roughly 2%, the U.S. Census Bureau said Thursday.

The once-a-decade national count put Texas’ official population at 29,145,505 after it gained the most residents of any state in the last decade, earning two additional congressional seats. In a post-count analysis using survey results from households, the bureau estimated that the count for people living in Texas households — a slightly smaller population than the total population — failed to find more than half a million residents. That’s the equivalent of missing the entire populations of Lubbock, Laredo and then some.

https://www.texastribune.org/2022/05...as-undercount/

[Steely Dan]

^ that same report said IL was undercounted as well, by roughly 250K people.

Which, if true, means that IL didn't actually shrink in population last decade.

And likely also means that Chicago is, in reality, back over 2.8M people again!

Not that the official census 2020 numbers will be changed or anything.


Here are the states with the largest under and overcounts (by %) from census 2020 according to the report.

Overcounts:
Hawaii +6.8%
Delaware +5.5%
Rhode Island +5.1%
Minnesota +3.8%
New York +3.4%
Utah +2.6%
Massachusetts +2.2%
Ohio +1.5%


Undercounts:
Texas -1.9%
Illinois -2.0%
Florida -3.5%
Mississippi -4.1%
Tennessee -4.8%
Arkansas -5.0

Source: https://www.npr.org/2022/05/19/10998...ennessee-texas

[lio45]

Originally Posted by Yuri

Not sure if someone mentioned previously, but the US mean center of population barely moved. In fact, only 12 miles, the shortest distance since the first Census back in 1790:


Wikipedia

Those are the growth rates for the US and the four major regions between 1990 and 2020:

UNITED STATES ---- 7,35% ---- 9,71% --- 13,15%

Northeast ----------- 4,14% ---- 3,21% ---- 5,48%
Midwest ------------- 3,08% ---- 3,94% ---- 7,92%
South -------------- 10,22% --- 14,29% --- 17,31%
West ----------------- 9,23% --- 13,84% --- 19,72%

The West has been growing slower than South since the 2000's, for the first time in history. The Northeast went against the trend and grew faster in the 2010's than the 2000's. All that is contributing to the slowdown.

And as the West has never been so close to the national average, it's not inconceivable that for the first time ever, the mean centre won't move any inch westwards, only southwards, specially now as it's already very close to Texan eastern border.

Ironically, it seems the "western" California that was the one pushing the centre southwards back for good part of the 20th century. Now it's the "southern" Texas that is the main force making the centre still keep moving westwards.

Actually, the entire Sunbelt has been dragging the center southwards. There’s a very clear inflection point (invention of air conditioning) in the trajectory of the center; for the first 150 years of the Republic, Americans were only moving due West.

[JoninATX]

Originally Posted by Steely Dan

^ that same report said IL was undercounted as well, by roughly 250K people.

Which, if true, means that IL didn't actually shrink in population last decade.

And likely also means that Chicago is, in reality, back over 2.8M people again!

Not that the official census 2020 numbers will be changed or anything.


Here are the states with the largest under and overcounts (by %) from census 2020 according to the report.

Overcounts:
Hawaii +6.8%
Delaware +5.5%
Rhode Island +5.1%
Minnesota +3.8%
New York +3.4%
Utah +2.6%
Massachusetts +2.2%
Ohio +1.5%


Undercounts:
Texas -1.9%
Illinois -2.0%
Florida -3.5%
Mississippi -4.1%
Tennessee -4.8%
Arkansas -5.0

Source: https://www.npr.org/2022/05/19/10998...ennessee-texas

That's good for Illinois, I wonder how many other states are truly undercounted.

[Yuri]

Originally Posted by lio45

Actually, the entire Sunbelt has been dragging the center southwards. There’s a very clear inflection point (invention of air conditioning) in the trajectory of the center; for the first 150 years of the Republic, Americans were only moving due West.

Indeed, the whole Sunbelt was responsible to push the curve south. What I meant was till recently California, and more specifically, Southern California grew a lot. So aside pulling the point west, it also contribute to pull it south as well, being a big centre of gravity.

Now, Southern California is growing less than national average and it's not contributing to push the point west, let alone southwards. It's Texas, a southern state, that aside pulling the point south, also contributes to pull it westwards, as it's still located west from the point.

In any case, it's nice to see the point barely moving. A sign "the North" growth is getting closer to the national average.

[jd3189]

Growth in Florida and Georgia/ North Carolina should also be pulling it south.

[Yuri]

Originally Posted by jd3189

Growth in Florida and Georgia/ North Carolina should also be pulling it south.

Yeah, but they are South pulling it southwards. Texas is mostly thought as South and it's helping to keep the point moving westwards as California is failing.

Westward movement is there since the US inception whereas the southward started relatively recently. We're very close, however, to see this point stop moving westwards. That would be interesting.

[BnaBreaker]

Originally Posted by Steely Dan

^ that same report said IL was undercounted as well, by roughly 250K people.

Which, if true, means that IL didn't actually shrink in population last decade.

And likely also means that Chicago is, in reality, back over 2.8M people again!

Not that the official census 2020 numbers will be changed or anything.


Here are the states with the largest under and overcounts (by %) from census 2020 according to the report.

Overcounts:
Hawaii +6.8%
Delaware +5.5%
Rhode Island +5.1%
Minnesota +3.8%
New York +3.4%
Utah +2.6%
Massachusetts +2.2%
Ohio +1.5%


Undercounts:
Texas -1.9%
Illinois -2.0%
Florida -3.5%
Mississippi -4.1%
Tennessee -4.8%
Arkansas -5.0

Source: https://www.npr.org/2022/05/19/10998...ennessee-texas

So in reality Tennessee likely has more like 7.3 million people, which would put it right around Arizona for 14th place potentially... this also likely explains the seemingly bizarre, sudden, and difficult to comprehend supposed population decline in Davidson County.

[accord1999]

Originally Posted by Steely Dan

Undercounts:
Texas -1.9%
Illinois -2.0%
Florida -3.5%
Mississippi -4.1%
Tennessee -4.8%
Arkansas -5.0

The undercounts likely cost Texas, Florida and Tennessee each 1 seat in the House.

[JoninATX]

Originally Posted by BnaBreaker

So in reality Tennessee likely has more like 7.3 million people, which would put it right around Arizona for 14th place potentially... this also likely explains the seemingly bizarre, sudden, and difficult to comprehend supposed population decline in Davidson County.

Yep, and the current Texas population would be around 31 million.

[iheartthed]

I understand undercounts, but how do they "overcount" if they aren't using estimation methods?

[galleyfox]

Originally Posted by iheartthed

I understand undercounts, but how do they "overcount" if they aren't using estimation methods?

Second homes, student housing, double counting.

If you see something like Florida with an undercount and New York and Ohio with an overcount, there’s probably some primary-secondary residences messing up the counts.

[Crawford]

Right, but this is circular reasoning. The count is the count. The estimates are imputation. The overcount/undercount is reconciling the enumerated count with the imputation.

Long story short, Census likely does a crappy job counting and imputing, given the wild discrepancies.

[Yuri]

That might be a silly question, but I wonder with all the cell phones, bureaucratic data, wouldn't be possible to have a population count even more accurate than the one in the tradicional census?

[iheartthed]

Originally Posted by galleyfox

Second homes, student housing, double counting.

If you see something like Florida with an undercount and New York and Ohio with an overcount, there’s probably some primary-secondary residences messing up the counts.

But the census form is tied to the residence, not the person.

[lio45]

Originally Posted by accord1999

The undercounts likely cost Texas, Florida and Tennessee each 1 seat in the House.

Depending on the exact wording of the Constitution, it's possible that that's perfectly normal. In which case, too bad for these states.