Paper short abstract:

This papers reflects on some of the tensions which arise when anthropologists collaborate with data scientists in analysing social data. I attempt to rethink these divisions, locating these frictions in situated practice and proposing new ways for anthropologists to intervene in these settings.

Paper long abstract:

Increasing numbers of ethnographers are studying the situated practices of data scientists and programmers producing algorithms, visualizations and apps using digital data (e.g. Seaver 2017). However, when data scientists work with 'social' data, this often brings their work into direct conflict with the methods and political commitments of their anthropological interlocutors.

This papers reflects on some of the tensions and power asymmetries which arise when anthropologically inclined researchers collaborate with data scientists in analyzing social data. These tensions are often reflected on but explained away as the great divide between quantitative and qualitative methods, positivist and hermeneutic approaches, or normative and descriptive outputs. It is often taken for granted that anthropologists deal in stories, oppose quantification and confine themselves to local settings and that data scientists are un-reflexive and concerned with reducing complexity (Neff et al 2017).

This paper attempts to rethink this divide by locating these tensions, not in fundamental orientations, but in situated frictions and instances of boundary work in everyday practice, drawing on empirical material from a series of collaborations with programmers. I find that these tensions can be productively negotiated but this means rethinking our given roles in these settings (as ethicist, communicator or detached critic) and potentially taking up the tools and terminology of data analysts, albeit in a more anthropological way. I argue that by intervening (Zuiderent-Jerak 2016) in these settings and causing frictions, we can learn more about these disciplinary tensions as well as exposing and testing our own normative commitments.