Big Social Data Approaches in Internet Studies: The Case of Twitter

Abstract

Well beyond Internet Studies itself, but arguably led by it to a considerable extent, there has been a turn towards computational methods in the study of social and communicative phenomena at large scale. This “computational turn” has commonly been described as a turn towards “big data” or, more specifically, towards “big social data,” and it continues to drive the development of new research methodologies, approaches, and tools. Internet Studies has been an advocate of “big data” approaches, because the field connects several core disciplines that use “big data” methods – media, communication and cultural studies, the social sciences, and computer science. Equally, the major objects of research in Internet Studies – including platforms, search engines, mobile apps and devices, and Internet technologies and networks themselves – are key sources of “big data” on user interests, attitudes, and activities. Proponents of such approaches suggest that it is becoming possible to “study society with the Internet,” while others ask critical questions about which observations are privileged and which are discounted as the logic of “big data” influences research agendas. The early development and application of “big social data” research methods in Internet Studies, as well as critical interrogations of such approaches, focused especially on research into Twitter as a global social media platform. This is largely due to Twitter’s (initially) highly accessible application programming interface (API), which enabled the development of powerful research methods and the promise of large, sometimes real-time, datasets tracing patterns of user activity around specific themes and topics on the platform, as well as, by proxy, in wider society. Twitter’s tightening of API access serves as a reminder of the precarious nature of “big social data” research drawing on proprietary datasets, just as concerns about the use of social media data for the social profiling of individual users raise questions about research ethics and user privacy. The growing body of “big data” research drawing on Twitter as a data source has paradoxically also underlined the many limitations and blind spots of such approaches, as researchers drawing on publicly available API data struggle to place their findings in the context of a platform whose overall global shape is shrouded in considerably more mystery, due to Twitter, Inc.’s interest in keeping aspects of the platform and its user community commercial-in-confidence. The increased work in this field also highlights shortcomings in research training and publishing models, which need to be addressed to further develop “big social data” research. This chapter outlines the current state of the art in computationally driven Twitter research, using platform-specific research as a case study for the computational turn in Internet Studies. It will consider the opportunities and challenges inherent in this shift toward more data-driven research and outline the key needs for the discipline which have emerged to date. Even as Twitter’s own fortunes fluctuate, the experiences made in this branch of Internet Studies stand as a guide for broader developments in our field.