Where is the most densely populated square kilometre in the UK? (2023) | Hacker News Reader

Where is the most densely populated square kilometre in the UK? (2023)

statsmapsnpix.com

68 pointsfanf253 comments

Interesting to note that for off the reputation of the UK as being all densely populated we actually still have more area of peat bogs (9.4%) and moors and heathland (7.5%) as urban fabric (~5.5%):

https://figshare.com/articles/journal_contribution/A_Land_Co...

Edit: Direct link to a map: https://orda.shef.ac.uk/articles/dataset/A_Land_Cover_Atlas_...

Edit2: That pastures are the top use (28.7%) maybe explains why "England's green and pleasant land" is justified! (And I say that as a Scot!)

And just under 10% of UK land, 20k square km, is reserved and optimized for grouse hunting, far more than every other sport combined.

Yes, agreed, but is that a core issue? These sort of comments seem like a diversion into class warfare.

In Scotland at least, I don't know about the rest of the UK, the use of large amounts of land for sporting estates (grouse shooting, stalking deer) probably isn't so much of a class issue but more about it being part of a very very long history (~1000 years) of land ownership issues, clearances from land, access rights, opaque ownership structures....

Also these outdoor sports are hugely ecologically unsound - driven grouse shooting (shooters stay still and are driven towards them) requires large amount of land to be burned each year so that young plant shoots grow to feed the grouse and tracks bulldozed into the hills, deer stalking has increased deer numbers so much (combined with other things like natural carnivores long been hunted out of existence) that natural tree cover has been terribly reduced across a lot of the Highlands.

Having said that, things have got a lot better in the last couple of decades - with "Freedom to Roam" landowners have got a lot more sensible about allowing access and most support it with a lot running side businesses with accommodation, selling venison etc. Also there are some incredible projects like Trees for Life aiming to rewild the Highlands and the Scottish Government and other agencies are encouraging community ownership of land - which can only be good. I should also say that there are few very wealthy people helping out in these areas as well.

So it's slow progress but I feel that things are definitely going in the right direction!

>> natural carnivores long been hunted out of existence

That is a trend seen in island nations. Japan and the UK both wiped out the large predators. Some speculate that when a country owns and entire landmass that the one government/culture gains control sufficient to enforce its will absolutely. Don't like bears? They can be wiped out with the passing of one law. Where landmasses are divided, no one government/culture has power. Take the US/Canada landmass. Countless US animals (ie wolves) only survive in the US from refuge populations in Canada. Had the US culture owned the entire landmass, as in the UK, then wolves would likely have been eliminated.

The lasts bears and lynxes were killed in Britain a long time before there were anything like unified political structures - indeed possibly even before the Angles (English) and Scots even arrived on the island. The last wolf was killed much later - usually given as the 17th century but even that is before Scotland and England were politically unified.

Class warfare is a core issue if anything is.

These areas are not nature parks. They are modified/burned to support grouse. Natural predators are trapped and farm-raised animals introduced to be later shot. A hunting reserve is not a nature preserve.

I know that. But a property used for hunting, including the burns, is a whole lot more ecological than a parking lot.

I'll take that over more urbanised areas. The natural environment isn't upper class.

Random squares is absolutely not cheating -- if you really really want a grid, you can always draw one out from the square you found.

Indeed, not picking the grid that gives you the best answer might be considered cheating :). Unless you're explicitly going for OS grid reference square kilometres, which the first grid may have been -- but I see no evidence that the author did so.

And if you want to really not cheat, you probably need to consider rotating your grid too.

All that said, I concur that it probably doesn't change the result very much. I'm just slightly disappointed that the author thought that improving his method to get a more correct answer was "cheating".

I've seen someone do the most populated circle in the world, with varying radius, and it's very interesting https://www.reddit.com/r/dataisbeautiful/comments/y51w28/a_n... . Lots of comments say that it's not the most objective but I say it is, because of your comment.

Completely agree. This is too easy if you use a standard grid.

One of the great things about London is that almost all of the 63 1km squares they highlight here contain a decent size park or other green space

I'd be curious how many of those spaces were open to the public 100 years ago.

J. Draper uploaded a remarkably interesting video recently about railings and green space in London.

https://youtu.be/JHvKvOUSor4

I wonder if it would be possible to find an even more populated square km, if one started turning it.

It would have been better if they looked for the densest circle with an area of 1 square km. Although a square shaped area has its tangible benefits - you "feel" population density more when it’s on your own street obviously, and since square areas provide more information along a few axes (the ones containing the corners), you can get an idea of the population density in an area by placing a square’s corners on the street. A square is probably not the best shape for this, though.

The other issue with squares is of course why not rotate the square? It seems silly to be bound to east/west and north/sout

You certainly would.

Just think how insanely unlikely it would be for this not to be the case.

One could also increase the density by allowing non square shapes or squares not fixed on the grid and so on. A more general approach would always lead to slightly higher scores.

> non square shapes

Perhaps something akin to a Hilbert curve that snakes around the whole country encompassing every individual person’s footprints but otherwise covers almost-zero area.

Sure, but thats a trivial solution. Exploring convex shapes however could lead to some interesting insights.

Yes, a grid system like H3 https://www.uber.com/en-NL/blog/h3/ is is closer to a circle in shape. One of the reasons for Uber to use H3 is bacause density maps look more natural in hexagonal shapes. It would be interesting to see the answer for H3 hexes.

There are a lot of nice things about H3. At work it's our default tool for partitioning and indexing of horizontal position data.

Are you using it for non geospatial data?

No sorry, just geospatial (horizontal position on the surface of the Earth).

It should be theoretically possible to re-develop it for other spaces (it's "just" a hexagon tiling), but the H3 team put a lot of work into the details and you'd be taking on a big effort redoing that for a different space.

> Here are some images from Street View - and recall that this is what about 25,000 people per sq km density looks like in London. This is half the density of the highest density places in Paris, Barcelona or New York, and about a quarter of the highest densities globally (found in cities like Manila, Cairo and Dhaka).

Huh -- high-density London looks like.. Queens, NYC?

NYC has a very large variation of density, from skyscrapers packed near each other to relaxed areas of 2-3-story buildings. Interestingly, it's not just packed Manhattan vs deep two-story Brooklyn. Manhattan is very uneven and has both 100-story buildings and 3-story buildings maybe a couple miles apart, and it's not downtown vs outskirts, but the general urban mix.

I remember reading somewhere that Los Angeles metro has higher population density than NYC metro.

I think the trick is the "metro." I've never been to LA, but have lived in NYC proper and now out in the greater "metro" and this does seem plausible to me, depending on how the metro is defined. If you look at a satellite map of LA, everything is squeezed into the valleys, but west and north of NYC city limits at least there's a lot more room to spread out (LI is more constrained) so that even though it's dense compared to, say, Omaha, I can well believe it's less dense than the LA basin and surrounding valleys.

I live about 13 miles west of the city as the crow flies, and its already suburban and not overly dense here, and just get more so west of here, still firmly within the metro.

I am trying to understand if the source data also provides the population weighted centroid. I guess it has to.no way to calculate that.

It would be interesting to know what tools were used for running the calculation.

Sheffield's Park Hill or Kelvin Flats, Leeds' Quarry Hill Flats, or Glasgow's Red Road Flats, might have challenged the spot-density at one point.

https://youtu.be/GELhFB-f8hQ?feature=shared

Is Bow E3 - Wiley - Topic what you meant to link?

Absolutely. The single most lyrically gifted grime MC hails from the single most densely populated area of the UK - Bayesians rejoice

I wonder what the most dense cubed kilometre is, in the world.

My loosely informed bet is on Hong Kong, but I'll be glad to be shown wrong

Could also be Singapore, Beijing or Tokyo.

I wonder what the most dense cubed kilometre is, in the world.

My loosely informed bet is on Hong Kong, but I'll be glad to be shown wrong

Before I even clocked on this I knew it would be in Tower Hamlets, famous for being the most densely populated council in the UK...

I wonder what a quadtree based approach would look like. I mean... I'd almost guarantee you're going to wind up in Tower Hamlets again. So the question of how much it costs to optimize for so little is there. But still, interesting alternative method for someone to gnaw on if they wanted.

How i would do it:

1. Create a table with quad tiles of appropriate size/level as Parquet. Put it in s3 and map it as a table in AWS Athena (Trino)

2. Put the census data in s3/Athena.

3. Join using Trino geospatial functions. https://trino.io/docs/current/functions/geospatial.html

The result for all quad tiles will be there within a couple of minutes.

Why though? We're talking about much less than a million geothings here. Any laptop can process this much data at interactive speeds.

True.

0. Look at data size.

Now I wonder what would be the most densely populated square kilometer in the world instead. I would imagine something like in Hong Kong.

A linked article says it's somewhere in Dhaka, Bangladesh, at "close to 200k per sq km".

https://theconversation.com/think-your-country-is-crowded-th...

I missed that, thank you.

I think so - or Singapore. Maybe Mumbai. Once upon a time called last year maybe a sqkm of Gaza would sneak in.

Kowloon Walled City will forever hold the record, I think. Density of 1.9M per square km. (Before it was demolished of course)

Wasn't that just because it was tiny (0.026 square kilometre) and a population of 35,000? I think for a discussion about "most densely populated square kilometres" we should actually be talking about square kilometres?

Fair enough. That 0.026 sq.km. has more people living in it than the most densely populated full sq.km. in the UK though, which must count for something. I wonder what the population of the square kilometre centred on KWC is/was.

But the city was much smaller than 1km², so doesn't qualify

Probably lalbagh in dhaka. ~170,000 per sq km.

Once I've heard that Mong Kok area was the most densely populated neighbourhood in the world.

No need to guess, https://en.wikipedia.org/wiki/List_of_cities_proper_by_popul... puts Manilla at the top of the list.

I asked for the most densely populated square kilometer, not the most densely populated city. A city's population can be spread out evenly or concentrated in a few areas.

Reply on news.ycombinator.com