Comments on: Zipf’s law and world city populations

By: A System Collapse Framework for Societies | 1913 Intel

A System Collapse Framework for Societies | 1913 Intel — Tue, 25 Feb 2014 09:08:05 +0000

[...] fires, sandpile collapses, snow pile avalanches, earthquakes, financial market collapses, wealth, city size, serial killers, riots, terrorism, attacks within war and wars. 16. In financial mathematics, the [...]

By: Irish Banks : Arrears, Deposits, Bail-in and Interest Rate Editon | Brian M. Lucey

Sat, 26 Oct 2013 06:34:17 +0000

[...] of this is a power law. Many many things have been found to follow power laws – terrorism, population of cities, bibliometrics, income distribution… theres no reason to think Irish bank losses [...]

By: Doi

Doi — Wed, 30 Jan 2013 07:10:21 +0000

Thanks for sharing. I got the following error when plotting with the following code.

 qplot(log10(1:nrow(x)), log10(pop), data=x) + geom_text(hjust=ifelse(1:nrow(x) %% 2 == 1, 0, 1),
+ label=sprintf("%s",x$city),size=2,colour='darkblue')

Error: When _setting_ aesthetics, they may only take one value. Problems: hjust,label

By: Mathematics.. A discovery or an invention? « Tentative Conclusions

Mathematics.. A discovery or an invention? « Tentative Conclusions — Sat, 08 Dec 2012 20:02:15 +0000

[...] the next frequently used word ‘of’. It is mysterious because, this pattern is evident in the population of world cities as well. Share this:Like this:LikeBe the first to like [...]

By: Demos

Demos — Thu, 17 Sep 2009 18:24:13 +0000

Hi There – Very cool charts! Was wondering if I could use or more of them in a presentation about society and demographics in the 21st century. Please email me if that’s OK – They are very descriptive of some trends we are discussing in some of our lessons.

Many Thanks,
DKO

By: Bob Carpenter

Bob Carpenter — Mon, 15 Jun 2009 20:19:54 +0000

I love these “chi by eye” experiments. How would you reject something being a power law?

If you look at Chinese Restaurant Processes, or their generalizations, Pitman-Yor Processes, they have a nice “explanation” of why data such as these might follow power laws.

By: Brendan O'Connor

Brendan O'Connor — Tue, 02 Jun 2009 23:47:42 +0000

yeah they probably are honestly. i was originally trying to figure out if/how sorting samples from a lognormal becomes powerlaw … but got sidetracked into these plots which are just too interesting :)

By: John the Statistician

John the Statistician — Tue, 02 Jun 2009 19:43:41 +0000

Are you sure they aren’t lognormal?

By: brendano

brendano — Thu, 28 May 2009 22:09:15 +0000

ah, very interesting point. so they were on a logspace-linear curve in the great old empire but are now stuck being too big for the smaller little truncated nation they’re now the capital of…

By: Stavros Macrakis

Stavros Macrakis — Thu, 28 May 2009 22:00:53 +0000

ggplot is great, isn’t it!

About your city-size analysis, where the largest cities are larger than expected, I suspect that this happens when a country is “truncated” compared to its former colonial or imperial extent. Moscow, Vienna, London, Brussels, Berlin, Delhi would be examples of this. (You might expect it for Istanbul as well, but Ataturk moved the capital to Ankara….).

-s