Data is not just the stuff of social scientific method; it is the stuff of everyday life. The presence of digital data in an ever widening range of human relationships profoundly unsettles notions of expertise for both ethnographers and data scientists alike. This collection situates digital data in broader knowledge-production practices. It asks about the kinds of social worlds that data scientists are creating as the profession coalesces, and looks at the contemporary possibilities available to both ethnographers and their participants for knowing, formatting and intervening in the world. It shows what digital data is doing to the empirical methods that sustain claims to expertise, with a particular focus on implications for ethnography.
The contributors offer empirically grounded accounts of the cultures, infrastructures and epistemologies of data production, analysis and use. They examine the professionalisation of data science in a variety of national and transnational contexts. They look closely at specific data practices like archiving of environmental data, or claims-making about how software is produced. They also offer a glimpse into the new methodological and pedagogical possibilities for teaching and doing ethnography in a data-saturated world.
can be grouped together as ‘neo-Bourdieusian’ ( Beerli, 2017a : 785). Most notably, Bigo (2014) shows how security practices are increasingly dominated by data analysts with knowledge of computer systems who have the potential to create and manage population statistically. Since the publication of Bigo’s work, we can see how the practices of data science, artificial intelligence and machine learning have rooted themselves in sectors such as agriculture, health and education
” Practices, ’ in Salah , A. A. et al . (eds), Data Science for Migration and Mobility ( London : British Academy ), pp. 364 – 81 . Ayoubi , Z. and Saavedra , R. ( 2018 ) ‘ Refugee Livelihoods: New Actors, New Models ’, Forced
that for data scientists, as opposed to ‘regular’ mathematicians, such virtues could be achieved only through deep and frequent submersion 64 Ethnographies of data science in the messier, more practical business of building machines and solving real-world problems. As a consequence, my informants spent a great deal of time learning how to do new things. They diligently studied for their university seminars, read papers, took online courses, sought out mentors, browsed peers’ repositories on GitHub, participated in programming competitions and hackathons, surfed
instance, rather than the population register, counting Finland’s population could be based on the number of residents who own a mobile phone. However, existing methods for official statistics are not suited to the analysis and interpretation of this data. Data science and data scientists are increasingly identified as the discipline and profession necessary for realising the potential of these new data sources, which require skills and knowledge of analytic techniques not typically used in official statistics such as machine learning, algorithms and predictive modelling
working to provide businesses with ‘emotionally intelligent’ analysis of customer feedback, 84 Ethnographies of data science focusing on their production of a data service through flexible alignments of ‘black box’ cloud computing systems. I show how data science is strategically distributed and reassembled through this effort, and highlight the opportunities and risks that arise within an entrepreneurial entanglement of self, product and market. By placing today’s efforts against earlier attempts to develop a national data processing industry, I show how
and research practices, associated with 184 Experiments in/of data and ethnography computational and ethnographic approaches, come to rub closely off each other. Based on recounting, from the ethnographer’s point of view, a number of ‘collaborative moments’ at the awkward intersection of computational data science and ethnographic fieldwork, the chapter explores not so much the disciplinary commitments involved in this interdisciplinary encounter as the nature of that very collaborative relationship itself. Of crucial importance here is the fact that, in ways not
11 The other ninety per cent: thinking with data science, creating data studies – an interview with Joseph Dumit Joseph Dumit and Dawn Nafus Editor’s note: This is a jointly edited transcript of an interview with Joseph Dumit (professor of Science & Technology Studies and Anthropology) about the Data Studies undergraduate minor being designed at University of California at Davis. This programme began in late 2015, and is led jointly by Dumit and Duncan Temple Lang, director of the Data Science Initiative at UCD, professor of Statistics, and formerly of Bell Labs
methods of verification and accountability. Those who are more concerned about the political structures and effects of this new resource also talk of data as ‘exhaust’ – the byproduct of human interaction that needs to be both ‘captured’ by the analytic converter of data science and properly managed and governed to mitigate the dangers associated with ambiguous attribution, security, corporate monopoly and nefarious techniques of surveillance and control. Most recently, other social, political and ethical questions have arisen about the implications of automation and
practice of modelling unstructured data – or what I call here the challenge of ‘baseless data’ – and how this might reframe conversations about the relationship between ethnography and data science. The chapter unpacks how the modelling of found empirical data in the context of climate science begins with the issue of how to create a baseline. I then reflect on the way in which this relationship between data and baseline reaches its limits in the analysis of social phenomena, opening up a space for ethnography to enter in. Dwelling in this problem of the limit of