The object of mobile spatial data,the subject in mobile spatial research

Introduction

IBM (2013) has estimated that nearly two-thirds of the global population keeps a mobile device within arm’s-reach at all times. Each day, we reach for these devices upward of 150 times, generating over 5 petabytes of data. Clearly the data generated through mobile device use is “big,” but it is also personal. Each device is, nominally, controlled by an individual. While data is generated through both active use and passive monitoring (Exner et al., 2011), each datum refers—in some way—to an aspect of the life of the individual using the device. This glut of new nominally individual-level data has become the focus of much economic speculation by corporate interests and research by a variety of academic disciplines. Although corporations and academic researchers leverage data for different purposes and under different imperatives (Schroeder, 2014), both tend to focus their analyses on the data produced, rather than the subject producing said data.

This commentary argues against the too quick leap from individual to data point present in much work on mobile spatial data and for the reseating of the reflexive, self-eliciting individual data producer as an appropriate object of analysis and key empirical means of untangling the sociotechnical processes at the heart of subject formation in a world suffused with data. To make this argument, I first note a tendency in past and current research on mobile spatial data to leap from individual to data point. I suggest this reliance on natively digital methods fits within a larger computational turn within the social sciences. I then suggest that one means for pushing past this focus on the produced data over those producing it is through the reconceptualization of mobile spatial data as a form of spatial media. Following from and extending existing definitions of spatial media, I argue that it, and the qualitative, empirical work that has come with it, opens for consideration the individuals who still generate the data points organized, sorted, and analyzed by algorithms. Finally, I provide two brief examples of recent work which has bridged the gap between data produced and the reflexive, self-eliciting subject that produced said data and demonstrate how such work deepens our understanding of the data of everyday life, allowing researchers to write thick descriptions of culture and subject formation in a world always-already mediated by data. ¹

The shifting state of spatial information

The term spatial big data has come to refer to a variety of sociotechnical apparatuses and objects of study that make any precise definition difficult. Hahmann and Burghardt (2013) have suggested that upward of 60% of all new digital information contains a spatial component, making the data generated through mobile device use far from the only source of spatial big data. However, for the purposes of this commentary, mobile spatial data is the type of spatial big data under consideration.

For geographers and other cognate disciplines, studies of spatial big data have often aligned with other terms like “neogeography” (Turner, 2006) which situate such work in specific cultural and historical contexts. The story, as it is usually told, begins in 2005 with the launch of Google Earth. The release of the Google Maps API (Application Programming Interface ² ) in June of 2005 granted users ready access to large databases of base map information that had previously been the domain of cartographic professionals. In this telling, the spatial big data can be read as the specifically geospatial implementation of a broader lowering of barriers to content creation and the resulting explosion of purportedly user-generated content referred to as “Web 2.0” (O’Reilly, 2005). Some version of the above narrative can be found at the beginning of many research articles on spatial big data (see for example Baginski et al., 2014; Cinnamon and Schuurman, 2013; Crampton et al., 2013; and elsewhere). These cogent, brief histories serve as shibboleths by which researchers signal article focus and “tribal” belonging (Haklay, 2012).

While attesting to the vibrancy and importance of spatial big data research, these narratives have tended to overlook some underlying tensions regarding the nature and production of the data itself. As noted by Elwood (2008), while the technological form may be new, many of the questions facing research on new forms of spatial data run parallel to those discussed in the mid-90s around the topic of GIS & Society. Issues such as privacy (Curry, 1997) and technocracies of access (Obermeyer, 1995) are mirrored in more recent work on the nature of geospatial data (Crampton et al., 2014). Further, by beginning the era with the release of a corporately controlled API, such narratives leave absent the social history of spatial information (Barnes and Wilson, 2014). As researchers have begun to fill in these gaps, there has remained a comparative focus on the outcomes of actions—the data produced—over the producing actions.

Such research uses what Rogers (2009) has called natively digital methods: methods which “diagnose patterns of social change via the digital traces that can be gleaned via the Internet” as opposed to more traditional social scientific methods directly involving individuals (Bruns and Burgess, 2012: 160). This reliance on natively digital methods can be read as part of what Berry (2011) calls the “computational turn” in the social sciences. This turn accepts that the information captured by digital device use is meaningful and provides “rich information” (Long et al., 2012) that can be used as a stand-in for the traditional subject. Taken to an extreme, the turn can have profound effects on the type of knowledge produced as models and visualizations of ever-larger data sets transform into “higher form[s] of intelligence and knowledge” (boyd and Crawford, 2012: 663).

For mobile spatial data, this turn often involves the treating of mobile spatial data as evidence of phenomena, rather than as phenomena in and of themselves (Wilson, 2014: 2). Researchers have taken the location of Flickr photos (Hollenstein and Purves, 2010), Yelp reviews (Baginski et al., 2014), and a variety of other geospatially tagged social media (Graham, 2010; Zook and Graham, 2007) as evidence of some spatially important distribution that reveals and makes a claim upon the world. For example, keywords tagged to locations in Google Maps have been used to study the spatial distribution of attitudes toward religion (Shelton et al., 2012) and Foursquare check-ins have been used to “discover [a city’s] structure” (Livehoods, 2012 in Thatcher, 2014: 1768).

While the methods and analysis of these works are exemplary, and many acknowledge some limits to their conclusions, when mobile spatial data is taken to signal the presence of a phenomenon in time and space, there remains an epistemological leap from individual to data point: a leap from the individual and whatever motivations went into the creation of the data point—the humor, sarcasm, irony, and earnestness that accompany everyday life—and the purportedly quantitative facts the datum comes to signal. Undergirded by a resurgent ‘pseudopositivist’ approach to quantification and meaning (Wyly, 2014: 30), this approach is nothing new. Geodemographics, for example, are built upon the leap from the individuals creating data to the areal unit described, and this leap is similarly echoed in claims around current spatial big data (Dalton and Thatcher, 2015). In academic fields, social scientists have long analyzed data from one context as a means of uncovering truth in another (for example, the common substitution of factors in economics).

While important to understand both the empirical value of such studies as well as their intellectual lineages, in a world in which mobile spatial data is increasingly the data through which individuals know and are made known to themselves and others, through performative processes such as the construction of the “spatial self” and enacting “conspicuous mobility” (Schwartz and Hageoua, 2014; Wilson, 2012), it is also necessary to study how data producers understand themselves and the data they produce. I will now argue that a key means of bridging the gulf between individual and datum is through the understanding of mobile spatial data as mobile spatial media. Following work by Crampton (2009), Kinsley (2013), Leszczynski (2015a), and others, such a move helps reframe mobile spatial data as “sites of potential relations between individuals; persons and places; and people, technology, and space/place” (Leszczynski, 2015a: 729). I extend this work by arguing that spatial media allow for a reconceptualization of research into mobile spatial data which reseats the reflexive, self-eliciting subject as an object of research in a data rich world, before concluding with brief examples of how methodological approaches that have incorporated the reflexive subject complement and deepen studies of this data of everyday life.

Between a twitch and a wink through spatial media

While information use remains implicit in many research designs, rather than an explicit focus of said research (Bertel, 2013: 301), spatial big data researchers have increasingly examined the partiality of mobile spatial data. No longer simple, direct evidence of a chosen phenomenon on a discrete, Cartesian plane, researchers are positioning mobile spatial data within relational (Crampton et al., 2013) and transductive (Kitchin and Dodge, 2011) notions of space. Rather than an inscription of meaning into a datum as evidence of some preselected phenomena, such conceptualizations recognize that the meanings of mobile spatial data may be multiple, influenced by and influencing assemblages of social, economic, and temporal factors that go into the production of space and identity.

In such research, mobile spatial data is understood to recursively shape and be shaped by larger processes of social and spatial production. A key means of pushing beyond hard subject–object divides and toward recursive, always-mediated understandings has been to approach some spatial data as a distinct form of spatial media. Constituting the technological objects that produce, store, and analyze data as well as the data produced (Crampton, 2009), spatial media consists of “digital networked spatial hardware/software objects and information artifacts” (Leszczynski, 2015a: 746).

Mobile spatial data is particularly well suited to be understood as a form of spatial media, because it is intrinsically generated through the daily “quotidian, vernacular uses” of mobile technologies (Leszczynski, 2014: 732). Mobile spatial data is created through tangled processes of mobile device use, the social norms that have formed around the use of said devices, and the data produced (or not) through said use. Mobile spatial data is personal data. It is the data that users generate when they tweet about their favorite (or least favorite) ice cream parlor, when they tag a friend in instagram, or—more passively—when their mobile phone pings a cell tower with their relative spatiotemporal location. As a move to push past understandings of hybrid duality between “real” and “virtual” spaces, to see the materiality of so-called virtual geographies (Kinsley, 2013), spatial media examines not only “the devices and information artifacts in question” but also the “practices with, through, and around these presences” (Leszczynski, 2015a: 736).

Extending these claims, I argue that such a move enables moving beyond the doing subject—the capturing and analysis of “activities and ‘doings’” that generate data—and toward the reflexive, self-eliciting subject that navigates and creates this form of spatial media in addition to analyses of the data produced (Ruppert et al., 2013: 34). The reflexive subject refers to how the data producer, the individual mobile device user, understands the data they produce, how they choose (and do not choose) to create data, and how this factors into their own (and others) understandings of themselves. Schwartz and Halegoua (2014: 1644) suggest such data practices reflect the emergence of a “spatial self,” constituted through the myriad of instances wherein individuals “document, archive, and display their experiences and/or mobility with space and place in order to represent or perform aspects of their identity to others.” The “spatial self” is not a complete representation of the individual, as users “carefully choose the places and times when these actions are broadcast” (Schwartz and Halegoua, 2014: 1648). This repeated, performative process of “making-known of one’s location is part of the making conspicuous of one’s mobility” (Wilson, 2012: 1270); it is a process of identity formation and inscription.

Spatial media involves an epistemological approach to mobile spatial data that includes not only the hardware and software used to create the datum and the datum itself, but also demands a critical, reflexive questioning of the technology, data, and social processes involved in their creation. This epistemology recognizes that data must be “imagined as data to exist and function as such”: it is never raw and always requires an interpretive base (Gitelman and Jackson, 2013: 3). The reflexive, self-eliciting subject is one means of bridging the interpretations generated through analysis of data and the attitudes and understandings of individuals who produce it. Interviews, participant observation of device usage, and deep ethnographies of hardware and software design are all means of bridging the gap between datum and individual, of moving from the doing subject to the reflexive one.

Following Geertz (1973), much as a boy rapidly moving his eyelids can only be said to be winking through always present “regimes of interpretation” (Boellstorff, 2013), two young women checking in to a restaurant on Facebook may have radically different intentions and motivations behind the identical interaction with their mobile device. From the perspective of the datum generated, these check-ins may look the same—the same format, accessed via the same API, and containing nearly identical digital information. Extracted from the individual, this data may then be tied together with the nearly 1500 data points per person data mining and marketing firms keep (Singer, 2012). This digital identity is the individual that processes of global capitalism can see (Thatcher et al., 2016). For example, one of the women may be marked as having an income well below the traditional clientele of the restaurant, her presence triggering a flag on irresponsible spending habits for her bank.

Rather than triggering on the fact that the woman has a lower income than the restaurant’s traditional clientele, an insight that can be leveraged from data mining and analysis, it is necessary to understand why the women has chosen that restaurant—perhaps she is celebrating a new, much higher paying job offer or is being taken out by a wealthy relative. The check-in may be motivated by this and intended to signal to the other party that she has arrived. Qualitative, empirical work, such as interviews and participant observation, are a key means of uncovering this richness. While corporate data mining operations, driven by a profit imperative, may have little interest in such nuance, for researchers this leap from individual to datum can have profound effects on the knowledge produced (Thatcher, 2014; Dalton and Thatcher, 2015). For example, while analysis of the check-ins would reveal the presence of the two women and this could be visualized and analyzed in a variety of ways, it would elide their intentions, what brought them to the restaurant, and why they chose to check-in. If we are to take seriously the “quotidian, vernacular uses” of mobile spatial devices and the data they produce (Leszczynski, 2015a: 732), then we must also take seriously the decisions that go into said use, the reflexive interpretations made by individuals as they do and do not make use of mobile spatial devices in their daily lives.

Certainly, the digital identity which emerges within broader systems of data extraction and analysis is a powerful tool for both researchers and capitalist investors, but the reflexive subject—replete with her own intentions, knowledge, and motivations—must also be an object of research. To move beyond this division between thick and thin interpretations, to rethink the social in the sociotechnical relations that produce culture, research in spatial media must recognize and analyze the reflexive interpretations of the end-users themselves, rather than simply the quantified, corporately controlled data they generate.

The data of everyday life

Mobile spatial data is an integral part of the emerging “quantified self-city-nation,” the entangled sociotechnical mesh through which individuals both come to and are made known, sorted, and (in)visible to themselves and society (Wilson, 2015). The systems involved, be they state data portals or podometers, are less about interpreting the existing world than in actively producing it (Kitchin et al., 2015). Again, this process is not new. The reduction of society to numbers for analysis is a long-standing function of modernity with statistics—the science of the state—being a core means of doing so (Foucault, 2008; Scott, 1998). However, representing the world as numbers works best when the world in question can be remade in the image of said numbers (Porter, 1995). The “quantified self-city-nation” promises to do so, producing a world that is inevitably and inexorably better, smarter, and safer.

But, at least in the case of mobile spatial data, the data leveraged is the data of everyday life. It is created with all of the intentions that go into walking the dog or tagging a friend in a photo. While ever larger data sets may produce statistically significant or seemingly generalizable insights, without depth they are at worst numbers “speaking for themselves” (Anderson, 2008) and at best the fallacy in which the world itself is as objectively precise as the numbers chosen to represent it (Whitehead, 1925). Mobile spatial data, generated at over 5 petabytes a day, is “big data” by almost any definition. While criticisms of “big data” have emerged in popular and academic presses (Carr, 2014; Taylor, 2015) and have focused on a variety of scales, such as individual and society (Edwards, 2014; Slee, 2014), critique requires not only theoretical sophistication, but also empirical depth.

This commentary has argued a key means of realizing empirical depth within spatial big data research is through the consideration of the reflexive, self-eliciting subject as an object of research. To do so I have drawn from and extended the work done on spatial media. Specifically, I have argued for the importance of the regimes of interpretation end-users create for the data they produce as a means of bridging the epistemological gap between individuals, their motivations, and their subject positions as they create data and the interpellated digital representation that emerges through aggregation and analysis of said data.

It is important to note that this is not a call for the cessation of research into the analysis of extremely large, heterogeneous sets of mobile spatial data. Nor is it to suggest that such studies cannot or do not acknowledge the variegated systems of data production. For example, Shelton et al.’s (2014) analysis of Hurricane Sandy tweets and Baginski et al.’s (2014) look at Yelp data both demonstrate the socioeconomic privileges inherent in the production of mobile spatial data. However, it is to suggest that in order to understand the specific motivations behind the production of data—and the decisions to not—it is necessary to understand the difference between a twitch and a wink, to write thick descriptions of the reflexive subject producing the data of everyday life (Geertz, 1973).

In the previous section, I described a hypothetical situation in which seemingly identical data points could hold drastically different meanings for the individuals creating them. In this conclusion, I want to call attention to two recent pieces of research on spatial big data which highlight the importance of the reflexive, self-eliciting subject for understanding data’s role in the world. Leszczynski (2015b) uses a small-scale survey to provide a greater context for Crawford’s (2014) concept of “surveillant anxiety.” Leszczynski highlights a specific desire to be able to “discern (be informed of, voiced in terms of concerns with transparency) and direct (maintain the contextual integrity of) flows of personal locational information,” rather than a generalized concern over its interception and analysis (Leszczynski, 2015b: 977, emphasis original). Through qualitative, empirical work with data producers, Leszczynski was able to deepen and contest Crawford’s previous claims. This was done not through a focus on the data produced, but through an engagement with how individuals understood the data they produced—whether they did so actively or not. Similarly, Dalton (2015) uses in person interviews with Google Maps developers to demonstrate the social norms and processes that come to shape the maps and visualizations made with new spatial big data technologies, offering otherwise impossible insight into how they may be transgressed and contested. Through a focus on how developers understood the technologies they produced, Dalton is able to call attention to the cracks between what said technologies represent and the individuals represented.

In a world suffused with mobile spatial data, the reflexive, self-eliciting subject helps bridge the gap between what data points represent and the understandings of those represented by them. Mobile spatial data is created through both active and passive quotidian practices wherein individuals both choose how to perform their “spatial selves” and are surveilled and represented as such. Qualitative and ethnographic methods that take as their focus the reflexive subject, the data producer, are a key means of untangling the processes that constitute the “quantified self-city-nation.” In the end, we must see mobile spatial data as part of the stories we tell ourselves about the data we give off to others (Geertz, 1973).

This commentary is a part of special theme on Spatial Big Data. To see a full list of all articles in this special theme, please click here: http://bds.sagepub.com/content/spatial-big-data.