Blended Data: Critiquing and Complementing Social Media Datasets, Big and Small

Abstract

Internet research, and especially social media research, has benefited from con-current factors, technological and analytical, that have enabled access to vast amounts of user data and content online. These trends have accompanied a prevalence of Big Data studies of online activity, as researchers gather datasets featuring millions of tweets, for instance–here, Big Data is a reference not solely to the size of datasets but to the wider practices and research cultures around large-scale and exhaustive (and often ongoing) capture of data from large groups,often (but not always) studied quantitatively (see Kitchin and Lauriaut 2014a;Crawford et al. 2014). However, the accessibility of“big social data”(Manovich 2012) for Internet studies research is not without its limitations and challenges,and while extensive datasets enable valuable research, combining them with small data can provide more rounded perspectives and encourage us to think more about what we are studying. Similarly, privileging the online-only or the quantitative analysis of social media activity may overlook or mask key practices and relevant participants not present within the datasets. We argue for a blended data model as a critique and complement for different social media datasets, drawing in part on our research into social movements and activists’ use (and non-use) of online technologies. Together, these approaches may overcome and negotiate the respective limits and challenges of social media data, both big and small.