Systems of Skills and Expert Knowledge: The Composition of Data Science (2007–2013)

• In Event: 161 - Section on Organizations, Occupation and Work Roundtable Session (one-hour).
  In Refereed Roundtable: Table 06. Social Capital and Networks II

Abstract

This article investigates a conceptually central but empirically understudied problem in the sociology of professions: Systems of knowledge. It proposes an operationalization that draws on established theories (Abbott 1988; Eyal 2013) to build novel tools for processing large amounts of textual descriptions (∼35k documents) of expert work. From the textual data it reveals social objects—the network structures of expert tasks based on the skill profiles they require—as observations for subsequent analyses of trajectories of professions in their social environment. The value of this approach is demonstrated in a comparative study of the recent formation of data science and three known cases. I show that by 2012 the knowledge structure underlying the rising profession of data science forms patterns similar to those of the well known and established legal profession. Both resemble images of interlocking fragmentation; abstract knowledge accommodates multiple specialized cores. Struggling professions (financial advisors and quantitative risk analysts), in turn, exhibit centralized knowledge structures. Qualitative examination of these patterns and their organizational environments are consistent with the existing literature: Legal expertise differs for individuals and organizations of different sizes, engineers rely more on technical than abstract knowledge, and a disconnect between creation of new knowledge and its application as well as substantively similar knowledge within and outside of institutional boundaries are observed as markers of weak professions. The analysis of data science shows how recent changes in the organizational environment allow previously distinct skill profiles to connect into a knowledge structure that grants independence from existing roles and hierarchies.

Author

• Philipp Soeren Brandt, Columbia University