Population Mountains

Firstly, wow! This is fantastic. And highly addictive.

But... hrmmm... the small town of Nowogródek Pomorski in north-western Poland (population 450) appears to have the tallest spikes in all of Europe. I wonder what that's about.

Actually, the anomalies (or "noise" as the article calls them) seem some of the most interesting parts of the map—it certainly makes it very easy to readily identify patterns in the GHSL in terms of consistent/predictable -vs- possibly unreliable data.

It's scary how populous the island of Java has become. Between Java and northern India, the way the map represents people, it looks like the land is overrun by people. Of course there are other high density areas but much of Africa and the Americas and obvs Siberia look like they have ample open lands without much people.

I hope economic growth and education kick in and allow some of these rapidly growing populations to stabilize. circa 100 million in the Pearl river delta --imagine a drought there.

Interactive version: https://pudding.cool/2018/10/city_3d/

Hold Ctrl while clicking/dragging to change 3D z-axis and rotation.

Makes it very clear what underlies the "this circle contains half the world's population": https://brilliantmaps.com/population-circle/

The visualisation is really nice!

Is there any text or CSV formatted version of the source data? When I went to the GHSL website [1], it says:

Format: the grid data are distributed as raster files in TIF format. The ZIP files contain raster files together with pyramids (i.e., TIF and OVR files).

I made MOBAC [2] profiles for every country in the world, but that's using administrative boundaries (country borders). I'd rather increase the zoom level in built-up areas automatically.

https://github.com/peterburk/mobacProfiles

[1] https://ghslsys.jrc.ec.europa.eu/faq.php

[2] https://mobac.sourceforge.io

This is a fun visualization but it seems to me it achieves the opposite of what it hopes to - it hides much of the population from view. Vast regions of the world with low population density appear empty - but they aren't empty, they simply have low population density. In aggregate these regions might have more population than the cities, but this visualization highlights dense regions. An individual pixel with a height of 60 in a city might be surrounded by 800 pixels with a height of 1. This is visually indistinguishable from 800 pixels with a height of 0, eliding an obvious difference in meaning.

I know a big part of this is getting the population data. But in the rare chance you do get your hands on geospatial data like this, Uber's Kepler tool is an easy way to start analyzing and visualizing the data with great results.

https://kepler.gl/#/demo

Doesn't '1,000,000 square meters' mean the same thing as '1 km²'?

I'm amazed at how flat Atlanta is compared to other cities. Even having lived there, I didn't realize the population density was that much lower than cities in the Northeast and on the West Coast.

This is great! I'm so excited for the wave of geo-spacial data demos using Mapbox's amazing API and their turf.js computation library that accompanies it.

I'd love to see this animated over time

Dang - this is a really cool viz.

This is addicting <3