Book Review: Algorithms for data science
Richard D. De Veaux, Nicholas R. de Veaux · Bulletin of the American Mathematical Society · 2017
When trying to define data science, the ancient Buddhist parable of the blind men and the elephant springs to mind.The entrepreneur, the academic researcher, and the university administrator approach the beast in turn and examine it, but report something different.The entrepreneur sees the training of a generation of data workers to deal with large data sets as a business opportunity.The university administrator hears the siren's call of new programs and students, while the academic researcher is still trying to figure out where the science is.An exact definition of data science remains elusive.As early as 1962 in The future of data analysis [15], John W. Tukey, a chemist turned topologist and finally statistician, wrote: "For a long time, I thought I was a statistician, interested in inferences from the particular to the general.But as I have watched mathematical statistics evolve, I have had cause to wonder and doubt. . .I have come to feel that my central interest is in data analysis."And later, "Data analysis, and the parts of statistics which adhere to it, must. . .take on the characteristics of science rather than those of mathematics-data analysis is intrinsically an empirical science."Later, in 1977, he published Exploratory data analysis [16], arguing that exploratory and confirmatory statistics are the two complimentary poles of statistics research.This sentiment was echoed years later in Leo Breiman's Statistical modeling: two cultures [2], where he chastised statisticians for focusing only on confirmatory analysis (which he referred to as the "data modeling culture") to the exclusion of exploratory analysis (which he called the "culture of algorithmic data modeling").He went on to note that other fields, notably machine learning and computer science, were rapidly filling this gap.In the same year, Bill Cleveland, then of AT&T Bell Labs, exhorted the statistics community similarly in Data science: an action plan for expanding the technical areas of the field of statistics [4], perhaps the first time the term "data science" appeared in print.According to the Data Science Association (DSA), "Data Science means the scientific study of the creation, validation and transformation of data to create meaning.A "Data Scientist" is a professional who uses scientific methods to liberate and create meaning from raw data-somebody who can play with data, spot trends and learn truths few others know."In the paper 50 years of data science, presented at the Tukey Centennial Workshop in 2015 [8], David Donoho claimed that to a statistician this sounds "an awful lot like what applied statisticians do."Indeed, the American Statistical Association (ASA) and Institute for Mathematical Statistics (IMS) at first reacted defensively with such articles as Aren't we data science [5] (column of ASA President Marie Davidian, AmStat News, July 2013), and Let us own data science [17] (IMS presidential address of Bin Yu, reprinted in the IMS Bulletin, October 2014).However, the ASA later came to the view that statistics is a necessary part of, but does not encompass what is currently understood as, the 2010