Median Center of Population for the United States: 1880 to 2020 (2021) | Hacker News Reader

Median Center of Population for the United States: 1880 to 2020 (2021)

census.gov

84 pointsLastNevadan78 comments

There is a lot of speculation here on what this means, so I looked it up. The median center is the point through which a north-south and an east-west line each divides the total population of the country in half. This is different from the "center of population", which is the balance point if you took an imaginary flat surface and put an identical weight on it for each person. These are both different from the geometric median.

This document explains how the census bureau computed the values and has equations. There are complications since the earth is a sphere (as far as the census bureau is concerned): https://www2.census.gov/programs-surveys/decennial/2010/prog...

> which is the balance point if you took an imaginary flat surface and put an identical weight on it for each person.

And with real weight for each person we'd see a center-point further south.

> This is different from the "center of population", which is the balance point

Does the difference only come from the earth curvature? I’m thinking if two lines each split the population in half, their unique intersection must also be the general balance point (for a flat surface anyway)

The balance point would rarely also be on the north-south, east-west intersection. The only time that would happen is if the distance of each person was also equal on either side of your meridian and your latitude.

In reality, the clumpiness of the populations on either side of your dividing lines are going to be different. You could have a dispersed population on the west side of the meridian, but a concentrated population far away on the east side of the meridian.

With center of gravity, adding a person far away from the center moves the center more than if you added them closer to the center.

Even in one dimension, the average and median don't have to be the same. In a list of numbers 1,2,3,4,100, the median is 3 but the average is 22. Same thing is done here, with same formulas to obtain them, just a two dimensional version (and some modifications to account for the meridians meeting at the poles, see the document linked by GP).

It will be interesting to see if this line starts moving back north if (when?) summers and heat in general continue to be this brutal. I'm a native Floridian and love it here, but I'll be damned if I haven't given thought to moving back somewhere a little more temperate after these past 2-3 months.

That's my guess. People will increasingly be moving away from the coasts due to flooding and storms and moving north to avoid the heat and other hazards (insects, diseases, etc) that are moving north and spreading.

Coastal real estate prices don't seem to match this prediction.

It does in some places (https://www.npr.org/2020/10/15/924239753/is-the-risk-of-sea-...) but I suspect it'll get much worse as climate change increases the frequency and intensity of storms and tax payers get more and more fed up with paying to repair the damage again and again in flood prone areas. (https://www.pewtrusts.org/en/research-and-analysis/articles/...).

California will likely be a hold-out much longer than other states. Earthquakes, wildfires, rolling blackouts, water shortages/restrictions, high unemployment, insane housing costs, high cost of living, and the worst traffic in the country haven't been enough to keep people away from California.

Florida is not affected by this yet, home prices have steadily increased except for a sell off during the pandemic.

https://fred.stlouisfed.org/series/FLSTHPI

Florida is going to be pretty turbulent in the nearish future I think. Theres already a looming insurance crisis as more and more providers pull out and/or raise rates. A serious hurricane could trigger some big price movements across the state I think.

You could make serious money (say.. 3-10x of the amount you put in) by shorting companies and buying puts if you strongly believe this.

I know there's a lot of reasons for this, but I think weather is one of the biggest. I wonder if that ever reverses from extreme heat. Is a full month of 110+ degrees having people reconsider Arizona?

I went the opposite direction of that line several years ago. It's been high 70's and low 80's all summer. Two low 90 days so far.

Last winter was mild as well.

So yeah, that could happen. Lots of water here as well.

Arizona is tax friendly to retirees allowing a more comfortable living on a fixed income. That can pay for a lot of air conditioning (or swamp coolers).

Air conditioning changed everything: https://www.energy.gov/articles/history-air-conditioning

> Engineer Henry Galson went on to develop a more compact, inexpensive version of the window air conditioner and set up production lines for several manufacturers. By 1947, 43,000 of these systems were sold -- and, for the first time, homeowners could enjoy air conditioning without having to make expensive upgrades.

> By the late 1960s, most new homes had central air conditioning, and window air conditioners were more affordable than ever, fueling population growth in hot-weather states like Arizona and Florida. Air conditioning is now in nearly 100 million American homes, representing 87 percent of all households, according to the Energy Information Administration.

People are way more tolerant of extreme heat compared to cold

Yes...up to a point. And there's the similar/related issue of extreme drought. If the wells and pipes run dry, the local population density tends to plummet.

Which is an artifact of a past age and foolish to bank on. Cold can be dangerous but we basically have what we need to survive at any low temperature in North America.

However, heat can't be combatted when you're outside in it with the same effectiveness. It reaches a point where the air temperature is simply lethal. It reaches a point where the paved ground burns whatever touches it. It degrades our medicines and damages our tools. We've already begun to experience these effects and it will only ever get worse. People should be moving away from hot areas, but instead they're running their AC 24/7 so they can keep pretending they made a good choice a little longer.

Until you get to wet bulb saturation extremes, you can survive heat as long as you have a supply of water. Cold temperatures, on the other hand, will kill you straight up.

Also, AC uses way less carbon than heat.

Heat pumps can work both ways.

We could reverse this statement, but replace "water" with "warm clothing". It's not like we haven't been ameliorating the effects of cold for ages.

> you can survive heat as long as you have a supply of water

Above ~55C (131F) that water will burn your tongue and throat. Even at 45C water will be uncomfortably warm.

What are you talking about?

A majority of the world's population is into drinking tea and other beverages significantly hotter than 55C

No, you're saying that like those extreme highs are distant specters. They already happen and are part of our immediate future. At which point, the human body cannot survive by evaporating water. Drinking cold water is an active measure which frankly is not available to all people, especially many of those most vulnerable to these fatal temperatures. When the WBT passes a certain level it will kill you straight up. Cold, not so much. People live in Antarctica. And although we will get storms and more extreme weather, the planet is getting hotter.

>> we basically have what we need to survive at any low temperature in North America

Senior citizens don't have what they need to survive shoveling snow at any temperature in North America.

Really cold seasons are cold around the clock... in the neotropical summer many people do go out running between 5AM and 7AM because it's fine to be outside when it's 75°-85°F outside.

At subtropical latitudes a low elevation desert climate with high temperatures overnight is very much the exception, not the rule. Likewise, it's unusual to have cold bright sunny seasons because few of us live in the Andes:

The hot subtropics are often bright even when it's muggy. This 24/7 makes one (even without going out jogging) feel less cooped up inside than the double whammy of low annual insolation gloom in so many cold latitudes when it's cold (even without having to walk in nor shovel snow).

A recent Atlantic article says, no. The county where Phoenix is, is still the fastest growing county in the USA, for multiple years running.

Most of the year the weather is good, apparently. Throw in air conditioning de rigeur, and cheap housing (due to no artificial supply restriction), and I guess you have the formula for population growth. Perhaps some politics are part of it too. Total taxes are half that of California. Probably not the #1 tax friendly state, but certainly up there.

Dewpoint and relative humidity are huge factors in how temperature is felt. 110F in Arizona might still feel more comfortable than 90F in eastern regions.

https://en.wikipedia.org/wiki/100th_meridian_west

>In the United States, this meridian roughly marks the boundary between the semi-arid climate in the west and the humid continental and humid subtropical climates in the east and is used as shorthand to refer to that arid-humid boundary.

I don't know. I was just in Northwest Arkansas with 100°F and 99% humidity and then a day later in Phoenix Arizona with 100°F and ≈30% humidity (at midnight!) and it felt similarly oppressive...

> Northwest Arkansas with 100°F and 99% humidity

People like to throw numbers like 90°F at 90%+ humidity. It doesn't work like this. If you look at the hourly data for Bentonville, AR (Northwest Arkansas) on the hottest day of the month [1] you see that by the time temperature got to 101°F humidity was at 32%. It is sticky as hell, but I guess as a number 30% doesn't sound impressive enough.

[1] https://www.wunderground.com/history/daily/us/ar/bentonville...

100°F and 99% humidity is a dew point of 100. The highest recorded dew point ever on earth is 95.

No, I do not believe that you were just in Northwest Arkansas with 100°F and 99% humidity.

I always feel hotter with the low humidity. It's stickier and possibly more uncomfortable if you aren't used to it with high humidity. But even low 90s with very low humidity feels like inside an oven to me. I was working outside all day yesterday and it was about 95 with with 60-70% humidity and it felt great with a slight breeze.

You were dehydrated.

Do we need a contemporary reason for it? People have been migrating away from the northeast since the early 1800s, and that is still going on today.

Looks like the 2020 line is just east of Chicago and just north of the SFBA.

The line can't move West or South too quickly as every additional minute represents so many people shifting from one side of the median to the other.

Not just movement east west or north south but also births, deaths, immigration, and emigration.

Also of note, the quadrants aren’t balanced. You could have 1x in the northwest, 2x in the northeast, 2x in the southwest, and 1x in the southeast.

Follows the shape of the gulf coast from northwest FL to Mexico approximately which would track the southwesterly flow of people

A couple interesting things here: first, what was that loop circa 1890-1940? I can't think of any major trends that would cause a westward and northward shift of the population except for the industrialization and de-industrialization of the midwest, though in my mind, de-industrialization (and thus population shifts) only occured in the 1980s or so. Second, why did it shift so drastically in the 1970s? You would expect a massive shift in the 1840s and 1850s with the Oregon Trail and Manifest Destiny, but I can't think of any shift in US history that would cause that large. Third, I'm curious where this dot would be during the Revolution (probably somewhere in North Carolina, I would guess), or during Columbus.

1. a combination of the continuing growth of the West and the Great Migration https://en.wikipedia.org/wiki/Great_Migration_(African_Ameri...

Consider the Great Migration of African-Americans from the Jim Crow south to the industrialising Midwest. 1930-1940 is also the height of the Great Depression and the Dust Bowl, so people would be migrating from the Plains areas down to California for better opportunities. With regards to the big shift in the 70s, the first thing that comes to mind is widely-available air conditioning making the interior of the southwest more habitable

I'm not sure what median Center of population means. Mean would be the geographic center accounting for location. Is the median just the 50th percentile in latitude and longitude?

I would guess it means 25% live in each quadrant of the map here:

https://www2.census.gov/geo/maps/DC2020/PopCenter/CenterPop_...

Not quite: half are on the left, and half are on the right of the vertical line; and half are on the bottom and half on the top of the horizontal line.

Make a list of lat/long for every person, pick the middle entries?

Someone else posted this link: https://en.wikipedia.org/wiki/Geometric_median

It’s the point the that minimizes the distance to all the input points in Euclidean space.

I'm going to guess this isn't what they actually mean. It's more likely just the spot where an equal number of people live left and right as well as above and below it.

You might be right, but I was more curious about what a technically accurate reading of the statement would mean.

OK, so it is basically/usually the a real data point closest to the mean location.

For something like population, where there are few gaps in data points, you would expect them to be basically the same because there is probably someone within a few miles of the mean.

This means that an extra person in Hawaii impact the "median center of population" more than an extra person in San Francisco because the mean is part of the calculation, opposed to just a stack ranking. Is this correct?

I'm surprised to see a data point for the 1890 census given that much of the records were destroyed in a fire.

https://www.census.gov/history/www/genealogy/decennial_censu...

Mathematically rigorous definition of "median center of population":

https://www2.census.gov/geo/pdfs/reference/cenpop2020/COP202...

I wonder what the median distance from the ocean is and how that's changed over time.

Technically the spot on the opposite side of the earth is also a median right?

Nope. https://en.wikipedia.org/wiki/Geometric_median

Thanks. That's an elegant definition of median!

Yes because it's the intersection of a meridian and a parallel, which is two points anywhere except the poles.

Amazing how little distance it's moved. Maybe it should be obesity-weighted to show it better.

Oh wait, that would not help. Maybe that's why it hasn't moved much...

Does this include Puerto Rico?

I'm guessing it doesn't - as PR is unincorporated. However, census is used in all sorts of things aside from voting/representation, so who knows.

What does incorporation matter? It's American citizens on American soil.

Puerto Rico's population is around 3 million, less than a percent of the 331 million population of the US. I wouldn't expect more than a slight change to the graph either way.

It is quite far south so it would have a significant impact on the mean, although not the median

South, but potentially east a little too. (PR is further east than all but parts of Maine).

Right now most of the Bay Area is barely to the south of the median, so I suspect the 2030 update won't change much there. As far as east-west goes it looks like it's in the middle of Chicago right now (which also helps explain why 2010-2020 didn't move that far west).

I brought up the Bay Area in terms of the medium because someone moving from the Bay Area to Hawaii would have zero impact on the median population longitude. However, it would impact the mean population longitude

Oh, I wasn't referring to your other comment.

Just noting that since the Bay Area's pretty heavily populated, bringing in Puerto Rico (shifting the US median south by ~1.6 million people) takes the median from say Napa's latitude to, say Alameda's latitude (keeping in mind that there's also Louisville KY and other smaller non-SFBA towns in the same band of latitude).

The computation uses the 50 states (or 48 conterminous states for calculations made prior to 1960) and the District of Columbia.

Source: https://www2.census.gov/programs-surveys/decennial/2010/prog...

Hmm... The "Dwight D. Eisenhower National System of Interstate and Defense Highways"? Started around 1956.

https://en.wikipedia.org/wiki/Interstate_highway

America is making a run for the border.

Westward expansion never stopped.

I believe Washington DC was close to the median center at the founding of the US. I've often wondered about the effect on US government if the seat of the executive (white house), legislative (capitol), and judicial (supreme court) branches had to move to the median center after every census. We need something to reduce the influence of K street. Something to keep the lobbyists from getting too cozy would be nice.

> Something to keep the lobbyists from getting too cozy would be nice.

That doesn't make much sense to me. The firms that rent office space on K Street don't have power because of the location of their office... it's the other way around: K Street real estate is rented out by those firms because of the street's proximity to power.

Move the power center and the firms currently renting office space on K Street, DC, USA will instead rent out space on Blah Street, Middlepoint, USA.

More-over, I don't know if moving the capitol of the country to the median point as defined in this article makes any sense either. It's one of those "literally everyone loses" propositions because "median" isn't "modal". If anything, it could make sense for the capital to be located at the midpoint by travel time.

But it's all wildly impractical if you stop and think about. The amount of infrastructure alone would require decades of work. And states would have to surrender sovereignty over a big chunk of their (settled!) land. Etc.

The midpoint by travel point mattered more in an era where travel across the country was measured in weeks or months.

At this point the only people with difficult transport are those in rural communities that may lack air services, though to combat this somewhat Congress funds the Essential Air Service at hundreds of millions of dollars a year.

Good point. Again, the whole thing feels a bit silly. Not a huge fan of DC but not sure why you'd think moving the capitol would actually change anything about how politics works...

I think this "move the capitol" thing is pushed by people who think "then DC would be accountable to small town America because it would be in small town America!"

Which, LOL, no. That's not how it'd work at all. That small town would lose self rule, be developed into a mid-sized city, become completely divorced from its previous character, and nothing else would change other than building a big ass city in Small Town, IL for no particularly discernible reason.

Which sets aside the fact that Small Town America already has an absurdly disproportionate amount of representation per human inhabitant.

Midpoint by travel time isn't too far from how French department capitals had to be reachable in one day by horse.

Sure if you give us a senate vote

There's an easy solution. Constitutional amendments written before the 1960s (more or less) had no expiration dates... they can be ratified centuries later. And have been.

It just so happens that there is an outstanding constitutional amendment, which was ratified by 11 or 12 states... it's already partially there. And if someone were to finish ratifying it, it would add almost 6000 extra representatives to the House (after the next census).

For one, this means that they'd need to figure out something other than the Capitol building, which can't seat that many. The probable result is multiple buildings, which can be outside of DC, and teleconferencing.

For two, this is far too great of a number for the lobbyists to be able to effectively bribe. Even now, they struggle (and sometimes lose) when attempting to push legislation their way. Raising the number of legislators they have to canoodle by x16 will strain both budgets and expertise.

For three, it probably also breaks the stranglehold the two parties have on Congress for at least the next 20 or 30 years. Their party mechanisms aren't large or sophisticated enough for this. They already have trouble sometimes fielding candidates in every race.

And Congress can't cockblock this. It's completely out of their hands. If someone were to convince their state legislature to ratify this, it would be newsworthy and catch the attention of other states many of which would follow suit. If a Wyoming or an Iowa or New Mexico were to give it any attention at all, I think it would inevitably become ratified.

Reply on news.ycombinator.com