Power Laws, Pareto Distributions and Zipf's Law
In recent years, researchers have become increasingly interested in the study of so-called power-law distributions.
¶
Heavy tails everywhere
¶
City sizes, word frequencies, citation counts, wealth distributions: many quantities follow power laws or close cousins (Pareto, Zipf). Newman explains the math, the mechanisms that generate them, and the sloppy measurement that invents them where they do not exist.
¶
Preferential attachment and beyond
¶
Some power laws come from rich-get-richer dynamics. Others from optimization or mixtures of distributions. The paper teaches you to ask what process before what exponent.
¶
For builders
¶
If you work on networks, language, or markets, this is the reference that keeps log-log plots honest.