This is an old revision of the document!
As a field in relative infancy, social media analytics has been populated by a plethora of methodological approaches which, for the large part, utilise linguistic understandings of the data to elicit insights - for instance, see Lewis et al. (2011)1) on how information and misinformation pertaining to the UK riots of 2011 was propagated via Twitter. There is as yet little consensus around what might count as a robust methodology for handling social media data, and this makes it crucial for researchers in the field to outline the detail of their practical working with data in such a way as to make the research process as transparent and work towards establishing a set of methodologies, methods and techniques which might be applicable to further research in this area.
Several key (related) conceptual ideas have been mobilized in our work with the #notracist dataset: our attempt to take a practice-based view of Twitter (as opposed to a strictly content-oriented linguistic one), and our conceptualisation of Twitter as an element in (various) techno-cultural networks or digital assemblages.
Our work with the #notracist dataset explores a materialist, practice-based understanding of Twitter in terms of how it works, which depicts language (and its expression on platforms such as Twitter) as a thing that people do, rather than strictly focussing on the content of what is said.2)
Whilst there may be value in, for instance, identifying and classifying tweets according to a “denial of racism” taxonomy - see Billig (1988)3) and van Dijk (1992)4) for examples of this type of work - to go beyond the categorisation of linguistic content in tweets may require a different approach. Our Analyses & Findings reflect this orientation.
Pursuing an understanding of Twitter and tweeting as something that people do in their social lives and interactions, it becomes apposite to treat Twitter as a set of practices involving multiple techno-cultural5) elements which connect together to form a digital (race) assemblage - see Sharma (2013).6) By “digital assemblage”, we are referring to the connectivity of numerous facets and factors of the tweeting process as it is experienced by people - this includes:
Our investigations into the #notracist dataset considers Twitter (and tweeting) as an assemblage that is connected to an array of other assemblages in complex ways, rather than treat Twitter as an isolated environment. We posit that the impetus to explore Twitter and tweeting practices in terms of their role in people's wider lives will help expand the field of social media analytics beyond examining statistics, social network analysis or conceiving users as autonomous agents outside of technological affordances.
Furthermore, our research work rests on an understanding that we, as researchers, find ourselves involved in the creation and navigation through digital assemblages of our own creation (involving Software Tools such as Chorus and Textometrica). Our work and publications resulting from this project attempt to reflect this by outlining how we build our peculiar assemblages, how we work with them, and how they result in findings contributing to understanding the materiality of online racialized discourse.