Vibrant Data

So, I think I will post this to Cyborgology later, but I would appreciate some feedback before I do…it feels a little too stuck in my current manuscript (or maybe *I* am just too stuck in my current manuscript).

Is data “vibrant” in the new materialist sense? It may not materially vibrate in the way sound waves do, but in its interaction with other phenomena (especially other data), data does exhibit the liveliness new materialists attribute to all things. In fact, some data scientists use concepts of “vibrancy” to describe data’s post- and extra-human capacities to percieve, know, and act.

For example, in 2013 Intel released a video called “Vibrant Data.” The video begins by contrasting linear perception to networked perception, and arguing that data is “a kind of augmented intuition” that can overcome the limitations of linearly focused perception, which phenomenologist Alia Al-Saji calls “objectifying vision.” Focusing on the persistence of a single signal through time, linear perception overlooks resonances among signals. In other words, by tuning into the primary signal, linear perception tunes out this signal’s overtones. Or, when we treat our lives as linear paths of first-person perspective conscious intentionality, we can only relate to those whose paths directly cross ours. It’s difficult if not impossible to find people whose patterns of behavior are in synch with ours if our paths don’t directly intersect. For example, if I go to the campus coffee shop on Tuesdays and Thursdays, and a researcher with similar interests goes to the campus coffee shop on Mondays and Wednesdays, we won’t know that it might be a good idea to talk about our work over coffee because our behavioral patterns, though strongly resonant, do not directly intersect. Data tunes into these resonances, to behavioral patterns beyond the spectrum of first-person subjective intentionality. As the Intel video suggests, data can connect us to people with “similar interests” and “overlapping circles of friends” but whom we have not yet “crossed paths.”

Intel’s video illustrates this with a story about “Veronica.” Interestingly, the Intel video uses music as a vehicle for data-augmented sociality: Veronica “listens to music most of the day” and thinks of her life “like a soundtrack.” Vibrant data finds Veronica a new favorite band, gets her to their concert in a town several hundred miles away, and in the process connects her to friends new and old. Data–or rather, “Veronica’s data”–knows that she likes this band, that it has an upcoming show in the region, that she won’t want to drive there, that somebody can fly her there, and that at the show she’s likely to run into some high school friends she’s lost touch with. It infers all these things from Veronica’s established patterns of behavior: what music she listens to, her transportation habits, and so on. Veronica’s Data both perceives and acts on information that’s imperceptible to and unknowable from Veronica’s first-person subjective perspective. In this way, as the video’s voice-over tells us, vibrant data gives us access to “experiences, connections, and possibilities we can’t begin to imagine.” Because data can access and process distributed networks of information that are invisible to the Modern subject’s first-person linear gaze, it can bring us more in tune with ourselves, with our surrounding environment, and with one another. And this in-tune-ness, the warm resonance with which vibrant data enriches our life, is described with reference to the “effervescence in the air” at a rock show. The affective and aesthetic high we get from listening to music we like with people we like is Intel’s metaphor for vibrancy.

Though Intel describes vibrancy with metaphors of aesthetic pleasure, it seems to function more like a Marcusean performance principle for algorithmic data processing. More precisely, vibrancy is a pleasure principle for us, but a performance principle for our data. Veronica–and all of us for whom she’s the surrogate–has access to this “effervescence” because “vibrant data is hard at work.” In fact, the full narration is: “there’s an effervescence in the air, our vibrant data hard at work bringing us experiences, connections, and possibilities we can’t begin to imagine.” Mining all our noise for the most resonant signals buried in it, data performs vibrancy in order to nurture and enrich our lives. Sure, vibrant data has agency: Intel’s video makes “data” (or, “Veronica’s data”) the subject of sentences: it “notices,” “decides,” “buys,” “knows Veronica,” “takes a leap,” “suggests,” and so on. But is it granted this agency only to indenture it in service to us? Do we grant data vibrancy so we can extract surplus, erm, vibrancy, from its hard work? This is, after all, what happened with white women in the US: they were given access to wage labor so they can be exploited by capitalist patriarchy not just as unwaged domestic workers, but also as feminized wage laborers.

This story of vibrant data completely obscures the fact that “vibrant” data can and is used to make some people’s lives more precarious, to subject them to disempowering and immobilizing surveillance. Credit data and NSA data are just as “vibrant” as Intel’s data. The “Vibrant Data Project” is much more aware of data’s ambivalent political potential. That’s why they emphasize vibrancy as a method of “democratizing data.” In an interview with  TED blog, founder Eric Barlow argues that “a more vibrant data system…encourages more people to participate.” Democracy, here, means participation. This is a nearly textbook example of liberal democratic theory: participation, often in the form of having a “voice,” is both necessary and sufficient for enfranchisement.

However, as Jacques Ranciere has argued, the very data science that Barlow and his Vibrant Data Project collaborators appeal to has transformed participation and envoicement into post- (which is to say, anti-) democratic practices. That is, data science has made participatory envoicement the very means of de-democratization. To explain, here’s a quote from a blog post I wrote on the topic. According to Ranciere,


Data is “the conjunction of science and the media” which understands itself as “exhaustively presenting the people and its parts and bringing the count of those parts in line with the image of the whole” ([Disagreement] 103). Data isn’t treated as a symbol or signifier of the facts, but as a measurement of the facts themselves….Ever-advancing technology “is supposed to liberate the new community as a multiplicity of local rationalities and ethnic, sexual, religious, cultural, or aesthetic minorities” (104). Twitter, for example, supposedly gives voice and access to people who are otherwise closed out of corporate [mass] media….The (supposed) advantage of “data” is that it allows us to think that we’ve solved all problems of justice, that we live in a post-racial, post-feminist, classless society, in a flat and perfectly meritocratic world. It looks like everyone is included, that everyone has a voice and that their voices count. From this perspective, the only injustices are making false claims about exclusion, marginalization, and oppression (e.g., calling out sexism gets interpreted as itself sexist).



The story we tell ourselves about data, that it is a means to universal envoicement, this story is itself the mechanism of post-democratic disenfranchisement. Following Ranciere’s model, we could say that data cannot make oppression or exclusion as something that is legibly wrong. So, though increasing data’s “vibrancy” might strengthen post-democratic institutions and modes of govenrmentality, it does not ameliorate oppression so much as naturalize it.

When data scientists talk about data’s vibrancy, they’re using vibrancy as a metaphor for agency, either of data itself (as in the Intel video), or of “we the data” (as in the Vibrant Data Project), that leads to a more dynamic, participatory, interactive, indeed, “effervescent” life. This effervescence is the affective, aesthetic pleasure that emerges from inclusion and participation in society. It is the feeling of being alive, that is, of having one’s life supported and facilitated by hegemonic institutions. This “effervescence” might also be understood as what Cristina Beltran identifies as “a kind of beauty that is experienced as a form of visible certitude” or “proof that we have collectively moved beyond prejudice and inequality and now live in a ‘post-feminist’ and ‘postracial’ era with institutions that are now fundamentally fair and accessible”(137-8). That is, it’s the euphoria or effervescence of feeling like one lives and participates in a truly inclusive, democratic society. From this perspective, it’s pretty easy to see how vibrant data is quintessentially biopolitical: it’s the use of statistics to manage and optimize the “life” of the population, of the “we” who are data.

As a technology, biopolitics is politically ambivalent: it can be applied in reactionary and radical ways. The effect of its application depends on a lot of things, including the material-historical situation in which it is applied, and how two key variables are defined. These variables are “life” and the “population”: what kinds of living count as healthy, viable lives, and, given the material-historical situation, whose ways of living most easily fit that definition.

The concept of vibrant data naturalizes those variables–that is, it turns them into constants. The metaphor of “vibrancy” defines life as something that is flexible, resilient, and agentially interactive. For example, as Barlow puts it, “vibrant” things are “moving parts” that “influence one another.” Data is “vibrant” when and because it affects other things, like data and, eventually, behavior. This definition of “life” takes phenomenological life experience of the most privileged members of society as the universalized, generalized mode of life as such. It overlooks the fact that these very same technologies fix oppressed groups in cycles of, as Stephen Dillon puts it, “immobility”: “The neoliberal state requires the management, regulation, and immobilization of surplus or expendable populations” (118; emphasis mine). Data profiles characteristic of oppressed populations, like poor credit scores, poor standardized test scores, and prison records, can make it difficult to access things like internet and/or wireless service, student loans, transportation, housing, and a lot of other things one needs to participate in the economy, the digital public sphere, to have an effect on others, to be a “moving part” of society that “influences” other parts. Again, this is classic biopolitics: the vibrancy and vitality of what appears to be the whole of the population is supported by the immobility and social death of those whose styles of living cannot be brought in phase with normative/hegemonic vibrations.

Data–or rather, algorithmically processed big data–does not literally, materially vibrate or resonate. Data’s vibrancy is just a metaphor for its liveliness, for its ability to come alive in support of the lives of those of us who are included in the “we” of “we the data.” Vibrant data is one example of how new materialist ontologies support white supremacist, patriarchal political projects.