Sociological engagements with big data: Analytical and methodological challenges
The rise of big data represents a watershed moment for the social sciences. Not only are we faced with large and multifarious types of data (e.g. texts, geo location, time stamps, entire full-text archives, pictures), often very unstructured, and stemming from all sorts of sources and phenomena, we are also challenged in our theoretical underpinnings of what constitutes the social and how we can analyze it. We are witnessing the rise of methods that help to identify patterns and relations, and to reduce complexity. Tools and algorithms of computational linguistics, machine learning, and network analysis are challenging the traditional tool kits of social science methods that work with samples, independent observations, statistical significance or analysts’ privileged positions in local settings.
My talk highlights the analytical and methodological challenges big data poses to the social sciences, and in particular to sociology. I discuss challenges of data construction, models of data analysis, and data interpretation. Moreover, I argue for a sociological engagement with big data analytics. Social networking companies have fully analyzed at least our social behavior online and data journalism provides colorful, interactive insights, for example, into social inequalities based on large public data sets. Sociologists, I maintain, should engage rather than refrain from such analyses: this might entail to fight back data scientists’ interpretation of the social, to challenge what is done with our data, or to adopt more data visualization in social science publications to name just a few. Whichever focus it may be, such an engagement certainly requires a sociological involvement with the tools and algorithms that collect, clean, sort, split, classify, and visualize. Thus, I argue for sociological analyses of algorithms as well as sociological analyses with algorithmic tools. In sum, my aim of the talk is twofold: to show the relevance of sociological insights for big data analyses while stressing the need to expand our theoretical horizons and tools kits in response to some of the challenges of big data.
Prof. Sophie Mützel,PhD, is Assistant professor at the University of Lucerne, Faculty of Humanities and Social Sciences
Discussant: Stefan Priester, Research Fellow of the WZB Research Group Science Policy Studies