After a bit of a break over the Australian summer and then a busy start to 2021, I have finally returned to the blog. Thanks for reading in 2020, and I hope I can share some more interesting work and ideas in 2021.
To that end, some great publications have come out in recent months. I am currently reading Veronica Barassi’s book ‘Child/Data/Citizen,’ but I have also enjoyed the recent article by Stark and Crawford on art in the age of AI, and Pereira and Moreschi’s piece on the unexpected ways of seeing with computer vision. It’s great to see the emergence of a ‘field’ focused on critiquing, exploring, and playing around with digital systems and data. I’m also looking forward to reading Lauren Bridges new article in Big Data & Society on unbecoming the ‘good’ data subject.
So far, in the Materialising Data project I have focused on how data can be represented to users in order to improve and expand understandings of digital data. However, a recent article in Social Media + Society by Neumayer, Rossi and Struthers, has been useful in thinking about the challenges to social media research posed by the (in)visible nature of data on social media platforms.
In many ways, Neumayer and colleagues start with a similar premise to that of Dourish and Mazmanian’s work on information materialisation, which I have blogged about previously. Like Dourish and Mazmanian, they contend that data are ‘made’ or socially constructed, and therefore need to be understood as a ‘complex set of political, cultural and scientific practices’ (p.5), including the technologies, people, discourses and imaginaries associated with data. However, this article explores the effects of this on social research.
The article argues for the need to disrupt the traditional approach to understanding visibility and invisibility as a dichotomy. They introduce the concept of ‘quasi-visibility,’ which refers to data that might be hidden but can be revealed to particular people or via particular technologies and/or levels of access. This ‘spectrum’ of visibility highlights ‘the processual character of data invisibilities’ (p.1) and provides a new and generative way to consider how power is manifest on social media platforms. For example, many of us may believe we are sharing a private message with a friend via a social media platform, but little do we know this is visible to the social media company.
How social media companies make data visible has implications for social researchers and the kinds of questions they can ask of data and platforms. They use the example of the social graph leading social researchers to use network analysis in their methodology. In this way, the social media company shapes how and what can be investigated and therefore the types of concerns that filter into mainstream public discourse. This is not necessarily a new idea, but seeing it through the lens of data and its (in)visibility is useful. As they explain toward the end of the article, ‘turning our attention to the invisible might inform our understanding of power relations as well as the socio-technical materiality of social media data’ (p.6).
The article identifies four socio-technical dimensions that shape whether social media data is rendered visible or not. They are:
- People and intentionality – ‘people leave traces of data with the aim of making visible but also to obfuscate and hide;’
- Technologies and tools – ‘as collections of “fact” are observed and stored;’
- Accessibility and form – ‘data are made accessible in various forms.’ Burgess and Bruns refer to these as ‘regimes of access,’ which are determined by social media companies.
- Meaning and imaginaries – ‘data are believed to have the capacity to measure, represent or unveil social phenomena’ (p.5).
The article concludes with a list of questions that draw attention to the limitations of researching social media data. What’s required, Neumayer and colleagues argue, is a more critical approach to the data collected on social media platforms, including what is made visible and why, and whether we have the technological tools to take account of the (in)visibilities.
In considering the materialising data project, Neumayer and colleague’s work helps us to look beyond critique of digital platforms and toward a consideration of the technological tools required to make invisible data visible. First we need to know, which data is (in)visible to researchers and why? Then we can ask, what kinds of tools do we need to better understand how data is used on digital platforms? After all, new ways of materialising data open up new epistemologies of data as well.