The metaphors of big data

Deborah Lupton has posted a really insightful analysis of the metaphors associated with big data. Deborah explores the different terms used to describe big data. The piece focuses predominantly on liquid type metaphors, but also includes some discussion of by-product type terminology such as ‘breadcrumbs’. She suggests that the metaphors used reveal a great deal about how we view the data, how we understand it and what our fears/hopes might be. I think I’m probably guilty of overusing the drowning in data type metaphors – although I’m not sure of exactly what terms i’ve used without looking back. I suppose it is a rhetorical device I have used when I’m trying to conjure an image. Deborah’s post is really thought provoking. Clearly the language used is powerful and will shape how we understand the data and what we might get from them. I’m hoping that this will appear in the book that Deborah is working on on digital sociology.

I wonder if the prominence of liquid type metaphors might not just be a product of the way these data developments are viewed, but might also result from dominant metaphors that were already present in social theory. These dominant liquid metaphors were already very popular in social theory and have perhaps found a new home in the big data debates. It got me thinking back to my Phd thesis. I had a section in there about what I described as liquid,fluid, flow metaphors. This was the early 2000s. At the time Zygmunt Bauman’s Liquid Modernity was proving really popular, as was John Urry’s Global Complexity. These, along with other texts, were frequently using liquid/fluid/flow as metaphors for describing the social world. In my thesis I tried to get inside these broad visions by focusing on the digitalisation of music. I asked what is fluid about these technological and cultural developments (as Deborah also points out, I found fixities and material practices). These metaphors seemed to become popular as a means for describing the apparent unanchored and decentralised complexity of the contemporary social world – with information technologies, globalisation, advanced consumer capitalism and the like. These metaphors also provided a new vocabulary away from the waning rhetoric of postmodern positions. they facilitated the move to alternative theories of modernity, and this served to assist in the transition between theoretical trends. I wonder if these existing metaphors, which had already become the means and resource by which we were thinking about and describing the social world, simply moved across into these new debates.

Deborah’s post also asks for some reflection on the term ‘big data’. This is not a term I tend to use very much. As well as being a bit imprecise, I also think it is a term that will date quickly. I tend to use the term digital by-product data (I’ve posted about my module here). This term is not much better but it doesn’t carry the same baggage as big data. As well as pushing towards visions of volume, as Deborah points out, ‘big data’ is also a little tainted with corporate and political discourse. This is not a problem in itself, but big data seems to arrive with a sense of its innate value. It is tied into a utopian vision, bigger is better, in which the scale of data makes everything solvable, efficient and predictable. I think this is why I’ve tried to avoid it where possible. Of course, it’s still possible to use the term without sacrificing a critical position. It is a helpful term of convenience, generally people have a shared understanding. It can be used with some self-awareness, but Emma Uprichard has pointed out some of the limitations and problems we face when getting dragged along with the big data idea. I’ve found I’m not comfortable with the term, this is for the very reason that Deborah’s post highlights. I don’t like it when the metaphors used, in this case big (if that is a metaphor), narrow down and constrain the way we approach the phenomenon. Deborah’s post is really worth reading because, as she points out, the metaphors we use are powerful in the way we approach these new forms of data. So they need some consideration, particularly in the early stages of development and as we search around for a shared vocabulary to capture the things we are trying to describe.

Advertisement