American Scientist

To the Editors:

As Brian Hayes points out in " Cultures of Code " ( Computing Science , January–February), there is more to studying the interaction of humans with computers than the algorithm-oriented computer science that one learns in an introductory undergraduate course. Nevertheless, despite all the diverse computer-related cultures mentioned in the piece, Hayes failed to mention the emerging field of data science. In this discipline, unlike the other computer-related disciplines discussed, the computation is focused on the structure and sometimes-subtle statistical relationships within large data sets. Moreover, data science is further differentiated from other computer-related disciplines in its emphasis on visualization to help tease out subtle but important patterns. With this new field of data science becoming popular in data-rich areas, such as the biological sciences and various branches of finance, perhaps it should be added as a fourth culture to Hayes’s canon.

Dov Greenbaum, PhD
Mark Gerstein, PhD
Yale University
New Haven, CT

Mr. Hayes responds:

I agree with Drs. Greenbaum and Gerstein that the recent emergence of data science opens a fascinating new chapter in the story I told in my column on cultures of code. I didn’t include the subject in the column because I don’t know how this part of the story is going to turn out. The field is quite new—it seems the first data scientists invented themselves in 2008—and it is still searching for a home in the academic world. Will data science be absorbed into departments of statistics, become a branch of computer science, or establish itself as an autonomous discipline? Or will it turn out to be a fad that eventually disappears?