Data (and) Science(s): The Double Movement of Continuity and Experimentation in Expertise Production

• In Event: Unpacking Expertise - I: Data, the Digital, and Organizations

Abstract

Quantitative data analysis has a long history in the sciences. Yet, over the past decade, academic institutions have also recognized the supposedly novel field of "data science." Our existing understanding of sciences leads to expect that data science has split from established disciplines to work on continuously specializing issues. In contrast, data science has entered academia from the outside and experiments with novel problems. I use data science's puzzling rise to shed light on the broader tension between continuous specialization and creative experimentation as alternative practices of scientific inquiry. My analysis builds on a combination of Polanyi's "double movement" theory and Peircean pragmatism. This combination links science’s continuity to its institutionalization and data science’s challenge of scientific institutions to abduction, research practices geared toward experimentation. I develop this argument empirically in data science’s academic history, showing scholarly attempts that failed and popular attempts that succeeded but met academic resistance. I also reveal the micro-foundations of this double movement process. I assembled a unique dataset to compare the co-citation structure of ~4.6k publications of scholars assigned to data science programs to the structure of co-citation networks in established academic fields. My analysis and application of data science reveal the interplay of institutional research scripts and individual experimentation with data in expertise production.

Author

• Philipp Brandt, University of Mannheim