Participative epistemology in social data science: combining ethnography with computational and statistical approaches

Abstract

ABSTRACT In this paper, I introduce the notion of participative epistemology and discuss how it can contribute to make social data science more accountable. I do so by offering the case of a project where ethnographic, computational and sequence analysis methods have been used in combination. By presenting here in greater detail research design and pilot results of a project using professional networking data to understand the careers of IT industry analysts, I suggest a view on the collaboration between data science and social science as coordinated labour. The application of participative epistemology to social data science is articulated in three points: (1) a more tactical view on the partnerships with commercial data where shared value system is not a pre-requisite for coordinated knowledge production; (2) an appreciation for complementarities in perspective between phenomenological sociology, expertise in computer science associated to digitalisation and the narrative positivism linked with the use of statistics and (3) a view on social data science as contributing empirical sociology with new sensitizing concepts, taking ethnography to reflectively address its own presuppositions.