Statistics for Data Science

We say that the string is 'random' if there is no other representation of the string which is shorter than itself. But we will say that it is 'non-random' if there does exist such an abbreviated representation. ... In general, the shorter the possible representation... the less random... On this view we recognize science to be the search for algorithmic compressions. ... It is simplest to think of mathematics as the catalogue of all possible patterns. ... When viewed in this way, it is inevitable that the world is described by mathematics. ...In many ways the search for a Theory of Everything is a manifestation of a faith that this compression goes all the way down to the bedrock of reality...

There are two kinds of statistics, the kind you look up and the kind you make up.

Although its evolution in the United States differed markedly from that of applied mathematics, statistics, too, benefited from the presence of the emigres and from the overall war effort. After a protracted period of professional differentiation from the social scientists and from the social sciences, mathematical statisticians had formed their own society, the Institute of Mathematical Statistics (IMS), in 1935. By 1938, the IMS had also taken over responsibility for the Annals of Mathematical Statistics, a journal that had been founded in 1929 to serve the needs of the more mathematically and theoretically inclined statistical practitioners. Thus, when refugees like Neyman, William Feller, Mark Kac, and Abraham Wald took up positions in the United States at Berkeley, Brown, Cornell, and Columbia, respectively, they were able to participate in a young, but viable, community of mathematical statisticians.