Rock Stars of Big Data? The Standardization of Expertise and Implications for Diversity in Analytics

• In Event: 445 - Section on Science, Knowledge, and Technology Paper Session. The Sociology of Big Data: Knowledge, Technology, Security and Privacy

Abstract

Big Data Analytics (BDA) is on the rise, yet “data science” lacks diversity like many other STEM fields. I argue that the role of technological infrastructures in shaping and maintaining structures of exclusion needs more attention, especially to the extent that technologies standardize and de-personalize particular realms of expertise. To support this argument, I present a case study in which I utilize Star’s concepts of “boundary objects” and standardization (Star 2010) to analyze publicly available online content from two recent professional BDA conferences. The traces of these conferences reveal a moment in which BDA assumes two distinct forms simultaneously. As some members of the community note the need for greater diversity among data scientists and raise notable questions about the Big Data enterprise, others attempt to standardize BDA by offering new technological infrastructures for managing and analyzing data. These technologies purport to “democratize” access to BDA by moving many technical decisions backstage and offering a point-and-click interface that requires little training or effort to use. While this may allow more participation in BDA, the assumptions and biases that shaped the behind-the-scenes decisions become increasingly invisible and difficult to challenge. The “rock star” Big Data experts—who are unlikely to be female, Black, or Hispanic/Latino—disappear even though their perspectives, assumptions, and experiences shaped the technology. Given the tendency of new infrastructures to be built upon the bases of previous infrastructures (Star 2010), the effects of the standardization of particular forms of expertise into technological infrastructure are likely long-lasting.

Author

• Margaret Willis, Boston College