This special issue brings together a number of articles focusing on the practical side of data science and data analysis. While there has been much hand-wringing about the rise of data science and its relationship to statistics, we avoid all such discussion here. Instead, the articles address the many aspects of day-to-day analytical work that are almost absent from conventional statistics literature and curriculum. Anyone who has ever taken wild-caught data through the full process of analysis knows that “statistics,” in the strict sense of fitting models and doing inference, is but one small part of the process. The remainder of the process accounts for a considerable share of the time and effort of data analysts, data scientists, and applied statisticians, but this is generally ignored in the statistics literature. We want to shine some light on these important areas.
We conceived this special issue with two audiences in mind: traditional statisticians, who may be most comfortable with journals, and the emerging, eclectic community of data scientists who communicate via preprints, twitter, and blogs. To reach both audiences, in addition to this special edition of The American Statistician, these articles also appear in preprint form in a PeerJ Collection (available at https://peerj.com/collections/50-practicaldatascistats/).
Our aim is to facilitate the transfer of tools and frameworks between industry and academia, between software engineering and statistics, and across different application domains. We hope that you find these articles thought-provoking, useful, and an interesting complement to the usual statistics fare.