Inside the Data Spectacle © The Author(s) 2014

15
Television & New Media 2015, Vol. 16(1) 37–51 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1527476414547774 tvnm.sagepub.com Article Inside the Data Spectacle Melissa Gregg 1 Abstract This paper focuses first on the scopophilic aspects of large scale data visualization— the fantasy of command and control through seeing—and places these in relation to key sites and conventions inside the tech industry. John Caldwell’s notion of “industrial reflexivity” provides a framework to explain the charismatic power and performative effects that attend representations of data as a visual spectacle. Drawing on twelve months of personal experience working for a large technology company, and observations from a number of relevant showcases, conferences, and events, I take a “production studies” approach to understand the forms of common sense produced in industry settings. I then offer two examples of data work understood as a new kind of “below the line” labor. Keywords Big data, data work, data sweat, below the line, scale, industry research Accounting for the spectacle of Big Data 1 entails understanding the aesthetic pleasure and visual allure of witnessing large data sets at scale. This paper identifies the scopo- philic tendency underwriting key sites and conventions inside the tech industry, which pivot on large scale data set visualization. I use John Caldwell’s (2008) notion of “industrial reflexivity” to explain the charismatic power and performative effects that attend representations of data as a visual spectacle, namely, the fantasy of command and control through seeing (Halpern 2014). Drawing on twelve months of personal experience working for a large technology company, and observations from a number of relevant showcases, conferences, and events, this “production studies” approach (Mayer et al. 2009) illustrates the forms of common sense produced in industry set- tings. 2 Due to the proprietary nature of high tech, few scholars have access to the points of ideological and intellectual transfer in which the promises of Big Data are 1 Intel Corporation, USA Corresponding Author: Melissa Gregg, Intel Corporation, JF-2, 2111 NE 25 th Ave, Hillsboro Or 97214, USA. Email: [email protected] 547774TVN XX X 10.1177/1527476414547774Television & New MediaGregg research-article 2014 at PENNSYLVANIA STATE UNIV on May 8, 2016 tvn.sagepub.com Downloaded from

Transcript of Inside the Data Spectacle © The Author(s) 2014

Television & New Media2015, Vol. 16(1) 37 –51© The Author(s) 2014

Reprints and permissions:sagepub.com/journalsPermissions.nav

DOI: 10.1177/1527476414547774tvnm.sagepub.com

Article

Inside the Data Spectacle

Melissa Gregg1

AbstractThis paper focuses first on the scopophilic aspects of large scale data visualization—the fantasy of command and control through seeing—and places these in relation to key sites and conventions inside the tech industry. John Caldwell’s notion of “industrial reflexivity” provides a framework to explain the charismatic power and performative effects that attend representations of data as a visual spectacle. Drawing on twelve months of personal experience working for a large technology company, and observations from a number of relevant showcases, conferences, and events, I take a “production studies” approach to understand the forms of common sense produced in industry settings. I then offer two examples of data work understood as a new kind of “below the line” labor.

KeywordsBig data, data work, data sweat, below the line, scale, industry research

Accounting for the spectacle of Big Data1 entails understanding the aesthetic pleasure and visual allure of witnessing large data sets at scale. This paper identifies the scopo-philic tendency underwriting key sites and conventions inside the tech industry, which pivot on large scale data set visualization. I use John Caldwell’s (2008) notion of “industrial reflexivity” to explain the charismatic power and performative effects that attend representations of data as a visual spectacle, namely, the fantasy of command and control through seeing (Halpern 2014). Drawing on twelve months of personal experience working for a large technology company, and observations from a number of relevant showcases, conferences, and events, this “production studies” approach (Mayer et al. 2009) illustrates the forms of common sense produced in industry set-tings.2 Due to the proprietary nature of high tech, few scholars have access to the points of ideological and intellectual transfer in which the promises of Big Data are

1Intel Corporation, USA

Corresponding Author:Melissa Gregg, Intel Corporation, JF-2, 2111 NE 25th Ave, Hillsboro Or 97214, USA. Email: [email protected]

547774 TVNXXX10.1177/1527476414547774Television & New MediaGreggresearch-article2014

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

38 Television & New Media 16(1)

actively debated and constructed. I offer instructive examples of this process, negotiat-ing the boundary of intellectual property restrictions and participant observation.3

The second objective of the paper is to theorize the labor of data. An important area of attention in the emerging data economy is to assess exactly how users’ online activ-ity involves them in profitable transactions, often without their knowledge (Scholz 2013). The analysis that follows adds nuance to this debate by identifying two instances of “below the line” labor (Mayer 2011) in the Big Data era. The first of these is the work of assembling the data spectacle, specifically the rhetorical work of the tech demo in selling the visions on display. This genre and its default evangelism are nor-mative features in the broader calendar of events for technology companies, large and small. Combined, they are a leading instance of what Caldwell calls critical industrial practice:

trade methods and conventions involving interpretive schemas (the “critical” dimension) that are deployed within specific institutional contexts and relationships (the “industrial” environment) when such activities are manifest during technical production tasks or professional interactions (labor and “practice”). (Caldwell 2008, 1)

Professional interactions in the high tech industry involve generating common-sense assumptions—of technology’s benefits, of technological progress as inherently good—a process that is pivotal to the broader experience of contemporary “data work.”4 Pursuing an analogy between the Hollywood locations that are Caldwell’s focus, and what is by now the rival center of mythologized cultural power in the United States, Silicon Valley, I use an example from a recent developer forum in San Francisco as an opportunity to unpack the ideological work of this type of industry event, one of many routine settings in which Big Data rhetoric launches and lands.5 These elite occasions for transferring insider knowledge operate as a flagpole running exercise for messages that will be sold to consumers later in the product cycle. Yet their distance from everyday users inevitably affects their ability to make appropriate judgments as to market desire and need. As such, tech events often pivot on a combination of self-aggrandizement and hot air recycling referred to in the industry as “eating your own dog food.”

The second aspect of “below the line” labor I attribute to Big Data is the work that data does on our behalf, with or without informed consent. Recent popular distrust of government agencies and technology companies colluding in the traffic of privileged information reflects the growing realization that labor in the new economy is as much a matter of non-human agency as it is the materiality of working bodies. After the algorithm has been implemented, sensors, screens, and recording tools require little human interference, even if the consequences of their scripts and commands only become known after deployment. The political economy of data exhaust (A. Williams 2013)—or what I will call, using a more organic metaphor, data sweat—requires deliberate strategies to overcome substantial power asymmetries (Brunton and Nissenbaum 2011). Informed by recent media studies documenting the environmental impact of machines that produce, harvest, and store Big Data (Gabrys 2011; Maxwell

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 39

and Miller 2012), the second part of this paper offers concepts that endorse responsible participation in a data economy. My hope is that these terms may assist in holding the purveyors of our data accountable for their actions.

In the move to a more “material” media studies (Gillespie et al. 2014), there has been a hesitancy to draw together humanistic thinking with notions of the non-human, a blockage that prevents a holistic account of labor in the digital conjuncture.6 Bringing these two aspects of data work together, I aim to demonstrate the combined relevance of humanities and social science methods in highlighting the ethical dimensions of technology innovation, which include the social consequences of data work at the level of the worker and his or her data. Given my position within the tech industry, my sense of the overall landscape for Big Data is perhaps more positive than others; it is certainly more optimistic than my reference to Debord’s Society of the Spectacle would imply. The objective of this article is to suggest that if the forms of representa-tion that commoditize our experience are today primarily visual (Halpern 2014), then television and new media scholars have a unique and urgent role.

Visual Pleasure and the Rhetoric of Data

The delight and comfort that can occur in the process of conceptualizing Big Data comes, at least partially, from witnessing the achievement of large data sets repre-sented at scale. The aesthetic pleasure summoned in these various constructions of data—from word clouds to heat maps or the color codes of quantification platforms—derives from their resolution of complex information through visual rhetoric (cf. Massumi 2005). “Beautiful data” is the result of a century of modernist thought dedi-cated to adjusting the ways we see, visualize, and manage information. As Halpern writes, in the Western tradition, vision “operates metaphorically as a term organizing how we know about and represent the world” (Halpern 2014, 19). It is

a metaphor for knowledge, and for the command over a world beyond or outside or subjective experience. To be seen by another, to see, to be objective, to survey, all these definitions apply in etymology and philosophy to the Latin root—videre. (Halpern 2014)

Sharing the same root as the word “evidence,” vision is the word that aligns truth and knowledge in different historical moments. In the case of Big Data visualization, it is “about making the inhuman, that which is beyond or outside sensory recognition, relatable to the human being . . . the formulation of an interaction between different scales and agents—human, network, global, non-human” (Halpern 2014, 18). The tech industry competes to provide this super-human insight via unique tools of data assembly. This explains why in corporate settings, the possibility of data visualization is regularly celebrated at the expense of considering the materiality of that which is processed. A recent company showcase provides a case in point.

At a demo booth illustrating the work of a research center dedicated to Big Data, onlookers were encouraged to watch, electrified, as synchronized TV screens dis-played dynamic images and patterns panning out from a point of origin. The effect of

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

40 Television & New Media 16(1)

this performance was doubtlessly impressive, even if, to a lay viewer, the morphing blobs of color brought to mind little more than the lava lamps and fashions of 1970’s disco. Engaging the spectator’s vision, simulating the experience of traversing (if not quite “tripping”) through data, the demo served the purpose of illustrating the vastness of the information being navigated. Yet when the presenter was asked, “What is the data set we are seeing?” it became clear that the data itself was fictive. There was no actual sample underwriting the demo, it was just a demo. The source of the data was irrelevant for a genre that only requires the indication of potential to achieve veracity. Like the trade rituals of film and video production, the tech demo exists within a wider ecology of “subjunctive” thinking that is the default mode of the developer forum: a means for “imagining—and showcasing—industrial possibilities on a liminal/corpo-rate stage” (Caldwell 2008, 105).

The affective properties of data visualization summoned by and through the demo bring to mind previous examples of the representing scale—the 1977 Ray and Charles Eames film, Powers of Ten, being the most familiar.7 In this sense, it was only fitting that a keynote speaker for the 2013 Association for Computing Machinery’s Computer Human Interaction (ACM SIG-CHI) conference in Paris was local sociologist Bruno Latour. The “expansive view” Latour chose to critique in his address [See Figure 1] drew from his previous writing on monadology (Latour et al. 2012). This work is informed by the ideas of Gabriel Tarde, and before him, Gottfried Leibniz, whose mathematical modeling questioned neat distinctions between individual and collective

Figure 1. Bruno Latour’s closing plenary, ACM SIG-CHI, Paris, 2013.ACM = Association for Computing Machinery’s; CHI = Computer Human Interaction.

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 41

phenomena. At a conference dominated by discussions about Big Data, Latour chal-lenged the congregation of industry and academic researchers, many of whom had relied on “the fallacy of the zoom” in their empirical reliance on data visualization. In Latour’s argument, a collective view provides no more accurate a representation than that of an individual—indeed, it is precisely the move to an expansive view that threat-ens accuracy and specificity.

Latour’s career-long investigations highlight the role played by tools in assembling vision. He questions the status and veracity of scale as a means of authorizing vision, and points to the labor left out of the frame, lens, or medium through which we view representations of reality. This approach acknowledges the selective nature of that which is “given” in what we think we see. The tool of assembly (the camera, say, or the algorithm) has agency in shaping sight toward certainties of apprehension. This recognition allows a degree of caution in thinking about Big Data when to do so means becoming unusually enamored with vision. It also suggests the relevance of aesthetics in explaining the role that visual pleasure plays in securing solace, excitement, and trust (Mulvey 1975).

The authority we attribute to scale is the result of historical accretion. According to Anna McCarthy (2006), initial definitions of scale rested on the musical sense of cap-turing a sequence of notes in order. Think of the gradually ascending tone structure of instruments we understand to be producing notes higher as opposed to lower in pitch. Like climbing a ladder, the series or progression implied in the idea of scale is a neat way to conceive relative order. We progress by degrees through positions that are taken to be naturally equidistant. Of the seventeenth-century thinkers McCarthy deter-mines as asserting this basic metaphysical hierarchy, Francis Bacon brought mathe-matical systematicity to the idea of scale. Central to this is an understanding of scale as proportion, which allows the significance of something to be observed “simply by comparing it to other things, without reference to external standards of judgment” (McCarthy 2006, 22). As a mode of reasoning, scale eventually stretched to influence not only practices of mapping geographical territory but also nascent ideas of political representation. Bearing resemblance to a thing—for example, a constituency—con-firmed the ability for something or someone to stand in place of and for others. This was also the period in which scale took on adjectival form. The consequences of this have proven resilient in the longer history of epistemology. Scale provides a “mecha-nism of translation, or mapping, which connects material things and their representa-tions in a precise, repeatable, and empirically known relationship which extends to the process of representation in thought” (McCarthy 2006, 23). Reason could move from the particular to the universal only as a result of these early articulations, which bestowed an obvious logic to graduating concepts of measure.

In McCarthy’s reading, scale “helps stabilize a necessarily murky dichotomy: the relationship between physical observation and mental speculation in inductive reason-ing.” From spatial representations of hierarchy (epitomized in the ladder) to dominant ideas of proportion (e.g., the map), a critical leap is necessary to join individual phe-nomena and broader conditions. Constructing the bridge between these two measures, “scale regularizes the process of knowledge production by implying that there is a

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

42 Television & New Media 16(1)

proportional relation between the datum, the definite axiom, and the general axiom” (McCarthy 2006, 24). The point here is that scale took on the function of reason through an induction, which constitutes a rhetorical maneuver. To summon the term scale is to mobilize “a thread of action and rhetoric actively connecting thought and thing, obser-vation and speculation” (McCarthy 2006, 25). The execution of this link, and the con-tinuum of empirical validity it suggests, is what we see playing out in tech demos today. Presenting data at scale invokes an epistemological claim in the mere act of display. It makes permanent what was once only plausible—a “cultural performance” of meaning that, while lacking a sound empirical referent, bears the hallmarks of the “instrumental and inductive perspective” favored in industry thinking (Caldwell 2008, 18).

Daniel Rosenberg (2013) offers another means by which to think historically about data’s rhetorical work. In previous centuries, he suggests, “datum” was understood as something given in an argument, something taken for granted. The obviousness of data, its taken-for-granted-ness, emanated from the Latin origin of the word, which in the singular means “gift,” or something that is “given.” In the domain of philosophy, religion, and mathematics, data was used throughout the seventeenth century to desig-nate that category of facts and principles that were beyond debate. It referred to things that were assumed, essential, and hence already known before a problem was intro-duced for discussion. Data contained the parameters for thinking, the foundation upon which later deductions would take place. Data is not, therefore, the same thing as fact. Data is something presumed prior to discussion, a framework creating the possibility for discussion. It therefore already contains judgments and decisions about what counts as a prior-ity (both priority and a priori share the same Latin root; priorities are taken from that which comes before). A data “set,” then, “is already interpreted by the fact that it is a set,” according to Travis D. Williams: “some elements are privileged by inclusion, while others are denied relevance through exclusion” (2003, 41). Like McCarthy’s etymology of scale, these details draw attention to the cultural specificity of reasoning. Even within the context of the English language, from previous usage, we see that

facts are ontological, evidence is epistemological, data is rhetorical. A datum may also be a fact, just as a fact may be evidence. But, from its first vernacular formulation, the existence of a datum has been independent of any consideration of corresponding ontological truth. (Rosenberg 2013, 18)

Rhetoric is a strategy of persuasion in the classical tradition. It is the art of convincing others the veracity and truth of something in spite of selective emphasis and exposure. So while we might continue to think of data as that which is given, as that which is regarded as bearing truth, we can see that the term’s shifting emphasis throughout his-tory removes considerations of partiality. Only recently did it become typical “to think of data as the result of an investigation rather than its premise” (T. D. Williams 2013, 33).

In the scripts tech workers perform during a demo, data’s power lies in the assump-tion that it is synonymous with fact. In the future-oriented mode of the genre, historic-ity is removed, and the benefits of the knowledge being assembled and transferred are

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 43

common sense. Taking a production studies approach, the further rhetorical effect at play in this process is the entrepreneurial imperative of the evangelist. If Caldwell warns of the dangers of industry-supplied PR in the Hollywood scene, and develops scrupulous methods to contextualize partisan spin, the digital optimism and venture-capital-directed pitching that constitutes the tech demo requires similar analytical pre-cision. It is not just the urgency and brevity of the encounter that illustrates the central role of rhetoric in this default industry ritual. In the developer forum, the selective showcasing of products and prototypes creates its own revelation, a preferred take on the best that a company currently has to offer. In these settings, all encounters have the character of a pitch (Gill 2011), right down to the questions of journalists and industry analysts whose career status rides in tandem with the quality of insights and scoops provided by a company’s star media performers. The hierarchy of access constituting these events means it is never simply a matter of reporting objectively from the show-case on offer but securing invitations to additional features and segments of uninter-rupted time with talent. Persuasion operates on a multitude of levels: in the data being presented, in the scripted lines of the worker out front of the demo, and in gaining access to what is a heavily orchestrated display of the present and future of computing. It continues into the press briefings, Twitter feeds, and column inches that construct the public’s apparently insatiable appetite for new media devices, technologies, and apps. In addition to the visual pleasure and power of data on display, then, the work involved in assembling and authorizing the spectacle taking place within the conven-tion center, tech campus, or downtown hotel is performed by a host of subsidiary workers acting after the fact, to one side, behind-the-scenes, and after hours.

Data Agents

If demo booths are a crucial site for the assembly and rhetorical illustration of Big Data’s commercial potential, the work that data does on our behalf—through data min-ing practices and other forms of network analysis—is an already established area of concern for media studies (e.g., Andrejevic 2013; Arvidsson 2011). From an industry perspective, the challenge posed by the data economy is less to do with limiting the scope of algorithmic surveillance as it is a race to define a profitable vocabulary for transactions that have the potential to bring new opportunities for connection, exchange, and wonder.8 If the prospect of data forming social relationships on our behalf brings untold risks, a business point of view sees infinite possibilities. The pro-liferation of music recommendation services (Seaver 2012) and online dating sites (Slater 2013) are just two of these convivial applications, in addition to the so-called sharing economy. With data as our agent, matching information with or without our direct involvement, algorithms create new matches, suggestions, and relationships that we are unable to achieve on our own. Data agents allow us to contemplate and revel in the possibilities afforded by strangers (Bezaitis 2013), whose profiles and tastes might anticipate or assuage our time-pressed needs. The very secrecy of online algorithmic sorting—the extent to which hook-up sites and platforms flourish through the partial revelation of identities and locations, for example—can foster collective social

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

44 Television & New Media 16(1)

practices that mainstream cultures may not wish to draw to light, presenting a boon for sexual and other minorities (Race, forthcoming).

My use of the term data agent thus refers to occasions in which the sorting, catego-rizing, and matching capabilities of data algorithms act as a highly competent append-age, a publicist, or even, to adopt some detective imagery, our shadow. In the world of Caldwell’s Hollywood, of course, agents have their own role. Agents act behind the scenes—their work happens to the side and in the background of stages upon which more visibly rewarding and profitable performances take place. Yet the agent’s work is essential in filtering a surfeit of information to a manageable and actionable set of options, matching available opportunities with potential investments. In the future already being built, the data we produce will be used to do something similar, that is, to work through algorithms to make decisions in our best interests, to sift out attractive or unsuitable options, and to favor encounters that accord with previously identified preferences. This is one way that data will entail new kinds of agency (if not an actu-ally existing, incorporated agency, such as the talent scout . . . although there may be merit in experimenting with this analogy too).

Decades ago, in The Presentation of Self in Everyday Life, Erving Goffman ([1959] 1973) relied on a similarly theatrical framework in his theory of region behavior. He divided social performances into two realms: the front region, which was deemed to be action on show to a public, and the back region, the site of relaxation and regenera-tion. Goffman suggested both regions host carefully cultivated performances that respond to cues elicited and interpreted in their respective settings. In the data society, a great deal of social work takes place off-stage, by non-human agents, as a result of processing choices engineered by computers. These programming decisions are made before any audience or user encounters the stage upon which communication later takes place. In orchestrating the setting for an encounter, algorithms and platforms are default editors for social messages. In assembling and choreographing the stage for digitally mediated performances, they also incorporate the work of key grip and set designer. An entire production studies lifeworld is employed in this complex infra-structure through which our data is assembled, and rendered visible and profitable. To recognize these layers thus requires engaging at multiple levels, part of a broader project of understanding the worth of “below the line” labor (Mayer 2011).

Data Sweat

Yet the idea of data agents still presumes a degree of distance between the individual and the information that circulates about an individual. It implies segregation as much as a process: I give my data to someone or something that can use it, hopefully to my advantage. Any number of events suggests the naivety of this aspiration, especially where there is a profit to be made. A more accurate way to think about our relation to data that avoids this gift economy is through the body. It is true, for example, that data may act like a shadow at times: our identifying data casts a shadow when we place ourselves in the glare of certain platforms or transactions. When recorded and pro-cessed at scale, data offers a rough outline of who we are and the form and function of

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 45

our digital projection for anyone motivated and literate enough to see. But this kind of analogy suggests we have some say in the interactions we choose to make, that we can predict, like the turning of the sun, the ways in which our data will be rendered visible and available. Instead of the visual metaphor of the shadow, then, we might consider an alternative and more visceral language to think past ocular-centric ideas of informa-tion sovereignty.

The idea of data sweat came to me in the course of giving a talk as a visiting speaker at a virus protection company in Taipei. The topic for discussion was data privacy and security, and as we were chatting, the air-conditioned building had a varied effect on the workers in attendance. Sitting in the crowded room, each person had their own way of dealing with the pre-typhoon heat, from fanning to slouching to wiping damp brows. Locals knew that any attempt to leave the building to walk the mid-afternoon streets would lead to gross discomfort. This contextual awareness led them to make all kinds of climate-dependent decisions, from choice of footwear (no heels) to transport (train or taxi), or just staying late at the office. One of the most enthusiastic audience mem-bers to introduce herself following my talk carried a tissue in hand to ameliorate her facial sweat, a taken-for-granted part of her daily ensemble.

Sweat is a characteristically human trait. It is a vital sign that our bodies are work-ing, even if cultural norms differ as to how much this expression should be public. In some cultures, for example, sweat can show enlightenment, possession, or commit-ment. It can just as easily suggest fear, anxiety, or arousal. Given this, sweat can appear when we may not want it. A whole industry of perfumes, deodorants, and other inno-vations now accommodates the need for disguise and masquerade in the process of maintaining social acceptability. Organic, corporeal phenomena such as sweat (but also microbes and genomes)9 illustrate the existence of data that is essential about us. This is data that speaks, albeit voicelessly, on our behalf. Sweat literalizes porosity: it seeps out at times and in contexts that we may wish it did not. It can be an annoyance or an accomplishment depending on the situation. But it is always a measure of our participation, our vitalism, and our presence in the social. Sweat leaves a trace of how we pass through the world and how we are touched by it in return. It is the classic means by which the body signals its capacity to “affect and be affected,” to use Spinoza’s terms. Understood this way, the labor we engage in as we exercise and exchange our data—especially in our efforts to clean up our image, present a hygienic picture, and make ourselves look good—is a kind of sweat equity for the digital econ-omy.10 It is a form of work we perform in the attempt to control what is ultimately out of our capacity.11

The current experience of Big Data is one in which powerful interests benefit from exploiting this lack of control. Turning the frame from one of personal sovereignty to data sweat gives us a better way of recognizing a rights-based contribution to this econ-omy; it describes the particular form of labor contributing to this common wealth (Hardt and Negri 2009). This is not labor that can be measured in terms of hours worked on the clock. To paraphrase Gordon Gekko: “data never sleeps.” Data work is beyond the mea-sure of “clock time,” and yet, to the extent that it generates profits that require compensa-tion, it requires us to think about value beyond measure. As Adkins (2009) argues,

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

46 Television & New Media 16(1)

While the break with the hegemony of clock time may lead to a break with certain kinds of measure—especially those forms which operate externally to entities—this break may also involve the emergence of new kinds of measure, specifically ones whose co-ordinates may emerge from entities themselves.

Data Exhaust

To move toward such an alternative way of thinking, I want to conclude by pushing the idea of data sweat to a plausible endpoint, through the notion of exhaust. This is not to signal exhaustion, since we have seen how data production and management takes place happily backstage, with or without our conscious effort. But rather, if data is a trail that we leave in our wake as a result of our encounters with the world and things, then this trail clearly has some undesirable effects. Within the tech industry, “data exhaust” or “tertiary data” names the value that our presence retains after a unique transaction (A Williams 2013). It is used to quantify the multiple applications that our digital identity provides beyond the gestures of an initial performance, to build busi-ness models based on the profits predicted from behavior cast by data. But exhaust is a term with further connotations, especially when thinking ecologically about the haz-ards posed by the massive computation of data on an increasingly fragile environment.

The clearest example of the environmental impact of Big Data is the investment in property and electricity now required by server farms that hold the world’s seemingly infinite packets of information. If data is the new oil, then data centers are the ports, wells, and tankers. The move to “cloud computing” is nothing if not a misnomer in this regard. Data that appears to be pushed to some higher, opaque place requires enor-mous physical infrastructure on the ground. To ignore these relationships, and the geopolitics they engender, is to perpetuate long-standing asymmetries in the experi-ence of computing (Pellow and Park 2002).

The further consequences of the data traffic moving between pipes and satellites across the globe include the logistical transfer, freight, assembly, and dis-assembly of always imminently redundant hardware (Rossiter 2014). Activists are documenting the human impact of this transport, manufacturing, and scavenging ecology, from the labor camps attached to Foxconn factories (Andrijasevic and Sacchetto 2013) to the Coltan mines of the Congo.12 As wealthy countries ship toxic e-waste back to the point of origin for disposal, the pleasures enjoyed through new social networks generate an international chain of service and manual labor. To evoke the legacy of an earlier moment of dystopic web theory, Big Data today translates to even bigger “data trash” (Kroker and Weinstein 1994).

Beyond the Sovereign Spectacle

An awareness of data exhaust invites us to take responsibility for the colonial legacy underwriting Silicon Valley mythology (Dourish and Mainwaring 2012)—the material conditions attached to the abstract philosophy of freedom through computing. If our

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 47

ideas of data are to remain wedded to the imaginary of prosthetics (something that is attached to, once it is taken from us), then ideas of sweat and exhaust may yet prove to have mobilizing potential. They can bring an assessment of environmental justice to bear upon the empowering mythologies emanating from Silicon Valley. The view I advocate in this paper, then, is that notions of personhood and sovereignty that per-petuate the fallacy that we can control our data will not assist in the cause of advancing an ethical data economy. We need terms that account for data’s agency in tandem with the human consequences of this new mode of production. Film and television studies provide a register to explain this double movement, in which the assembly of data and its capacity to act on our behalf each instantiate a form of “below the line” labor.

In his classic account of The Gift, Marcel Mauss ([1922] 1990) explains that noth-ing of value ever really comes for free. The forms of obligation that accompany a gift are social and pressing. They involve calculations of honor, status, and reciprocity. To offer a gift is to offer a part of oneself—the object is “never completely separated” from the instigator of the exchange. In a highly mediated economy, in which data is often traded without our knowledge, Mauss’s theory takes an interesting twist. If we are never fully aware of the context in which our data is given, the social bond that is formed lacks guidelines and nuance. The terms of obligation demanded of the giver and receiver remain compromised and unclear.

To date, Big Data has appeared as a gift for tech companies seeking to reinvent themselves from the triumphant years of desktop computing and lead the charge into a new market for software services, security, and storage. As this frenzy has taken place, we have lacked a human vision of rights in what is now regularly referred to as an “Internet of Things.” Television and new media studies have always acknowl-edged connections between the worlds of business, entertainment, and everyday life, and governance (Andrejevic 2004; Miller 2001; Ouellette and Hay 2008). And just as audience studies needed the insights of production studies to square the account, Big Data demands analyses that are attuned to both on-screen and behind-the-scenes components of digital life. This paper identifies a vital role for new media theory in encouraging better descriptions of data work. Applying media studies methods to Silicon Valley not only expands the reach and purchase of these legacies for a new moment but also creates a new set of political and ethical questions for the field. Writing from an industry position—from inside the data spectacle—I hope to encourage greater numbers of voices and actors to engage directly with those work-ing “below the line” in the data economy, to speak loudly in support of different and more inclusive casting choices and participants, and to drive different possibilities for computing and data processing from within. In the data industries of the future, a range of skills and literacies are going to be necessary to maintain just and fair opportunities for all. As I have shown, it is the rhetorical and visual effects of data compiled in the aggregate that television and new media studies are especially well placed to assess. The aura enacted in the performance of the data spectacle demands both theoretical precision and appropriate accountability. It requires new rights to be imagined and secured for the mass of individuals currently captured in—if not wholly captivated by—Big Data visions.

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

48 Television & New Media 16(1)

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

1. I use the capitalized proper noun throughout in recognition of the special issue this article joins. For a more specific discussion and critique of the Big Data conjunction and its pres-ent popularity, see the collection of papers assembled from research in the Intel Center for Social Computing in Maurer (forthcoming).

2. Writing this paper coincided with my first year as Principal Engineer in User Experience at Intel Labs, USA. As co-director of the Intel Science and Technology Center (ISTC) for Social Computing, my role is to work with academic partners across multiple universi-ties on five organizing themes: algorithmic living, creativity and collectivity, materialities of information, subjectivities of information, and information ecosystems. These topics provide a framework for collaborative research that guides industry professionals to better understand the social aspects of computing that may be overlooked in traditional engineer-ing approaches. This paper draws on observations and conversations at a range of ISTC and tech industry events in the United States, Europe, and Taiwan over a twelve-month period. Specific conversations are acknowledged where possible.

3. While my key reference for this kind of industrial reflexivity is Caldwell (2008), another inspiration for this paper is Georgina Born (2004), whose rigorous study of machinations within the BBC was a source of consolation throughout my first year at a leading technol-ogy company.

4. I am indebted to Katie Pine for this term and ongoing observations of how instruments for auditing, accountability, and measure affect the everyday experience of a range of work-ers, especially in the fields of health care and medical practice. See, for example, Pine and Mazmanian (2014).

5. Caldwell’s notion of production culture explains the behind-the-scenes labor underwrit-ing Hollywood’s primary position in the film and television industry. It also offers a use-ful frame for the unique configuration of cultural authority now emanating from Silicon Valley. Social anxieties currently attached to tech work in the Bay Area bear an interesting correlation to previous concerns about television. To name just a few, how each commu-nication technology (television vs. the Internet) creates a new industry for targeted adver-tising; the overinflated concentration of industry talent in one geographical area (LA vs. San Francisco); the celebrity status of key participants (screen stars vs. hackers), and their exceptionalism in the face of social norms; let alone the universalizing ideological aspira-tions of the industry as a whole, which, as a form of “soft power” in international trade and diplomacy, acts as an index of U.S. imperialism. Thanks to Jason Wilson for helpful conversations on these points.

6. Referencing the new materialism risks conflating specific traditions of thinking that encompass the actor-network theories and applications inspired primarily by the work of Bruno Latour, various strands of materialism understood through Deleuzian vitalism

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 49

(e.g., Braidotti 2013), German media theory traditions now most closely aligned with writers such as Parikka (2012), and object-oriented ontology (Harman 2002). In the Intel Science and Technology Center (ISTC) for Social Computing, the materiality of informa-tion theme has conducted research on auditing and measure that accompany the quantifica-tion of society (see Nafus, forthcoming); it also refers to the material practices of making, hacking, and repurposing that are accompanying the rise of consumer DIY electronics and maker culture. For another attempt to avoid binaristic thinking in labor theory, see Qiu et al. (2014).

7. See http://www.powersof10.com/film. Accessed June 15, 2014. 8. The somewhat discordant experience of intimacy produced through this novel combination

of global communications infrastructure, logistics, and system sorting is deftly captured in the Facebook slogan, “Ship Love” (Sloane 2014).

9. Thanks to Lana Swarz for prompting this idea.10. Thanks to Ken Anderson for the idea of “sweat equity,” and for many other forms of sup-

port as I wrote this article.11. Ellie Harmon takes this idea one step further to suggest that companies such as Facebook

are like the bacteria that live on our bodies and sweat. Personal communication, June 25, 2014.

12. See http://www.gongchao.org/en/frontpage for updates on Foxconn in particular. Accessed June 15, 2014. The Guardian has covered the ethics of Coltan mining for several years: see Taylor (2011) for a moving example. In January 2014, Intel CEO Brian Krzanich announced a new industry standard for sourcing “conflict free” minerals. See http://www.intel.com/content/www/us/en/corporate-responsibility/conflict-free-minerals.h-tml and related activism through the “Enough” project: http://www2.american-progress.org/t/1676/campaign.jsp?campaign_KEY=6265. Accessed June 15, 2014.

References

Adkins, Lisa. 2009. “Feminism after Measure.” Feminist Theory 10 (3): 323–39.Andrejevic, Mark. 2004. Reality TV: The Work of Being Watched. Lanham: Rowman &

Littlefield.Andrejevic, Mark. 2013. Infoglut: How Too Much Information Is Changing the Way We Think

and Know. New York: Routledge.Andrijasevic, Rutvica, and Devi Sacchetto. 2013. “China May Be Far Away but Foxconn Is on

Our Doorstep.” Open Democracy, June 5. http://www.opendemocracy.net/rutvica-andri-jasevic-devi-sacchetto/china-may-be-far-away-but-foxconn-is-on-our-doorstep (accessed August 16, 2013).

Arvidsson, Adam. 2011. “General Sentiment: How Value and Affect Converge in the Information Economy.” In Sociological Review Monograph Series: Measure and Value, edited by Lisa Adkins and Celia Lury, 39–59. London: Wiley-Blackwell.

Bezaitis, Maria. 2013. “The Surprising Need for Strangeness.” TED@Intel. http://www.ted.com/talks/maria_bezaitis_the_surprising_need_for_strangeness.html (accessed August 15, 2013).

Born, Georgina. 2004. Uncertain Vision: Birt, Dyke and the Reinvention of the BBC. London: Random House.

Braidotti, Rosi. 2013. The Posthuman. Cambridge: Polity.Brunton, Finn, and Helen Nissenbaum. 2011. “Vernacular Resistance to Data Collection

and Analysis: A Political Theory of Obfuscation.” First Monday 16 (5). http://dx.doi.org/10.5210/fm.v16i5.3493 (accessed June 8, 2014).

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

50 Television & New Media 16(1)

Caldwell, John T. 2008. Production Culture: Industrial Reflexivity and Critical Practice in Film and Television. Durham: Duke University Press.

Dourish, Paul, and Scott Mainwaring. 2012. “Ubicomp’s Colonial Impulse.” In Proceedings of ACM Conference in Ubiquitous Computing, 133–42. Pittsburgh, PA: Association for Computing Machinery.

Gabrys, Jennifer. 2011. Digital Rubbish: A Natural History of Electronics. Ann Arbor: University of Michigan Press.

Gill, Rosalind. 2011. “‘Life is a Pitch’: Managing the Self in New Media Work.” In Managing Media Work, edited by Mark Deuze, 249–62. Thousand Oaks: Sage.

Gillespie, Tarleton, Pablo J. Boczkowski, and Kirsten A. Foot. 2014. Media Technologies: Essays on Communication, Materiality, and Society. Cambridge: MIT Press.

Goffman, Erving. (1959) 1973. The Presentation of Self in Everyday Life. New York: Anchor Books.

Halpern, Orit. 2014. Beautiful Data: A History of Vision and Reason since 1945. Durham: Duke University Press.

Hardt, Michael, and Antonio Negri. 2009. Commonwealth. Cambridge: The Belknap Press of Harvard University Press.

Harman, Graham. 2002. Tool-Being: Heidegger and the Metaphysics of Objects. Chicago: Open Court.

Kroker, Arthur, and Michael A. Weinstein. 1994. Data Trash: The Theory of the Virtual Class. New York: St. Martin’s Press.

Latour, Bruno. 2013. “From Aggregation to Navigation: A Few Challenges for Social Theory.” Keynote address to the ACM SIG-CHI Conference, Paris, April.

Latour, Bruno, Pablo Jensen, Tommaso Venturini, Sébastian Grauwin, and Dominique Boullier. 2012. “The Whole Is Always Smaller than Its Parts: A Digital Test of Gabriel Tarde’s Monads.” British Journal of Sociology 63 (4): 591–615.

Massumi, Brian. 2005. “Fear (The Spectrum Said).” Positions 13 (1): 31–48.Maurer, Bill. Forthcoming. Big Data. Prickly Paradigm Press.Mauss, Marcel. (1922) 1990. The Gift: Forms and Functions of Exchange in Archaic Societies.

London: Routledge.Maxwell, Richard, and Toby Miller. 2012. Greening the Media. New York: Oxford University

Press.Mayer, Vicky. 2011. Below the Line: Producers and Production Studies in the New Television

Economy. Durham: Duke University Press.Mayer, Vicky, Miranda J. Banks, and John Thornton Caldwell. 2009. Production Studies:

Cultural Studies of Media Industries. London: Routledge.McCarthy, Anna. 2006. “From the Ordinary to the Concrete: Cultural Studies and the Politics

of Scale.” In Questions of Method in Cultural Studies, edited by Mimi White and James Schwoch, 21–53. Malden: Blackwell.

Miller, Toby, with Nitin Govil, John McMurria, and Richard Maxwell. 2001. Global Hollywood. London: British Film Institute.

Mulvey, Laura. 1975. “Visual Pleasure and Narrative Cinema.” Screen 16 (3): 6–18.Nafus, Dawn. Forthcoming. The Quantified Self. Cambridge: MIT Press.Ouellette, Laurie, and James Hay. 2008. Better Living through Reality TV: Television and Post-

welfare Citizenship. Malden: Blackwell.Parikka, Jussi. 2012. What Is Media Archeology? Cambridge: Polity.Pellow, David, and Lisa Sun-Hee Park. 2002. The Silicon Valley of Dreams: Environmental

Injustice, Immigrant Workers, and the High-Tech Global Economy. Cambridge: MIT Press.

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from

Gregg 51

Pine, Kathleen, and Melissa Mazmanian. 2014. “Institutional Logics of the EMR and the Problem of ‘Perfect’ but Inaccurate Accounts.” In Proceedings of ACM Conference on Computer Supported Cooperative Work, 283–294.

Qiu, Jack Linchuan, Melissa Gregg and Kate Crawford. 2014. Circuits of Labor: A Labor Theory of the iPhone Era. tripleC: Communication, Capitalism & Critique. Forthcoming.

Race, Kane. Forthcoming. “Party ‘n’ Play: Online Hook-Up Devices and the Emergence of PNP Practices among Gay Men.” Sexualities.

Rosenberg, Daniel. 2013. “Data before the Fact.” In Raw Data Is an Oxymoron, edited by Lisa Gitelman, 15–40. Cambridge: MIT Press.

Rossiter, Ned. 2014. “Logistical Worlds.” Cultural Studies Review 20 (1): 53–76.Scholz, Trebor. 2013. Digital Labor: The Internet as Playground and Factory. New York:

Routledge.Seaver, Nick. 2012. “Algorithmic Recommendations and Synaptic Functions.” Limn 2: Crowds

and Clouds, August 16. http://limn.it/algorithmic-recommendations-and-synaptic-func-tions (accessed August 5, 2014).

Slater, Dan. 2013. Love in the Time of Algorithms: What Technology Does to Meeting and Mating. London: Penguin Books.

Sloane, Garrett. 2014. “Mark Zuckerberg Gets Reflective as He Nears 30 Espouses Motto of ‘Ship Love’ at the f8 Conference.” Adweek, April 30. http://www.adweek.com/news/tech-nology/mark-zuckerberg-gets-reflective-he-nears-30-157394 (accessed August 5, 2014).

Taylor, Diane. 2011. “Congo Rape Victims Face Slavery in Gold and Mineral Mines.” The Guardian, September 2. http://www.theguardian.com/world/2011/sep/02/congo-women-face-slavery-mines (accessed August 5, 2014).

Williams, Alex. 2013. “The Power of Data Exhaust.” TechCrunch, May 26. http://techcrunch.com/2013/05/26/the-power-of-data-exhaust/ (accessed September 9, 2013).

Williams, Travis D. 2013. “Procrustean Marxism and Subjective Rigor: Early Modern Arithmetic and Its Readers.” In Raw Data Is an Oxymoron, edited by Lisa Gitelman, 41–59. Cambridge: MIT Press.

Author Biography

Melissa Gregg is a Principal Engineer and researcher at Intel Corporation. Her publications include Work’s Intimacy (Polity 2011), The Affect Theory Reader (co-edited with Gregory J. Seigworth, Duke 2010), Cultural Studies’ Affective Voices (Palgrave 2006), and Willunga Connects: A Baseline Study of Pre-NBN Willunga (2011).

at PENNSYLVANIA STATE UNIV on May 8, 2016tvn.sagepub.comDownloaded from