With thanks to Federico Foderaro for help with the MaxMSP coding.

Introduction

1Every day, our relationship with our data (collected facts and pieces of personal information), our bodies, and our perception of ourselves intertwines increasingly with the digital. Lupton (2019) explains that this is partly due to our pervasive use of digital technologies, that produce data that in turn influence our present and future experiences. Take, for example, how personalised advertising works to modify behaviour and attitudes (Zuboff 2019), or how constructed user profiles can influence people‘s ‘access to opportunities such as employment, travel, or credit, and even their choice of dating partners’ (Lupton et al. 2018: 649). Alternatively, popularised self-tracking technologies help us understand ourselves, engaging us emotionally and producing feelings like self-control, satisfaction, or disappointment (Lupton 2021). Yet, our relationship with our bodies and technology is not only one of entanglement and symbiosis, but also of embodiment, as Don Idhe (2002) exemplifies with how hearing aids and glasses assist us in our perception of the world. 

2This evolving relationship between our data, selves, and bodies is nourished by the quantification of daily practices. Mayer-Schönberger and Cukier (2013) describe this using the term datafication. This word encapsulates how all areas of life (like time, body temperature, sleeping, walking), and even qualitative aspects (such as spirituality, friendship, fashion, happiness) get coded, reduced to numbers, analysed, shared, and turned into capital (Van Dijck 2014; Bode and Kristensen 2016). The rhetoric surrounding datafication, big data, and the Quantified Self movement paints this new measurability as capable of unearthing not only self-knowledge but also societal understanding. In the words of Chris Anderson, editor in chief of technology and lifestyle magazine Wired (2008), ‘with enough data, the numbers speak for themselves.’ However, although data promise to bring us closer to understanding ourselves and society, datafication concurrently creates distance. Specifically, presenting information in a standardised numerical form depersonalises a phenomenon, as it distances the individual from the phenomenon represented and fosters the perception of objectivity and truth (Porter 2020). Boyd and Crawford surmise that ‘large data sets offer a higher form of intelligence and knowledge that can generate insights that were previously impossible, with the aura of truth, objectivity, and accuracy’ (Boyd and Crawford 2012: 663).

3However, faith in numbers and the conception of their objectivity is not new. In the 18th and 19th centuries, many scholars such as Florence Nightingale and Francis Galton used numbers (aka. statistics) to uncover the ‘laws of society’ (Manovich 2011). During that time, various fields, from psychology to biology, physics, and chemistry, similarly asserted that the social and natural world could be understood in simple terms. The trend towards statistics and reductionism through numbers traversed scientific fields and emerged in governance and state practices. Anthropologist James C. Scott (2008) explains that collecting and analysing citizen information has been practised for centuries, because it helped make society legible, and thus helped control it. The public health sector exemplifies this in that policy actors have used health statistics and population data to govern and regulate society for decades. As Mayes (2015: 23) explores in his book The Biopolitics of Lifestyle, population data and statistics help govern and regulate both individuals and the population, in that by making personal choices visible, ‘techniques of governance can be mobilised’.

4Beyond uses in public health and governance, our mass data generation fuels big data analytics, which promises value generation opportunities for various industries. We see this in the boom of social media during the last years, a business model underpinned by users continuously creating, engaging with, and responding to content (Lupton 2020). Digital platforms, today, commercialise aspects of our online and offline life, using social relations and networking apps as foundations for revenue. However, such value promises have not only ‘created an attitude of collecting data without a pre-defined purpose’ (Günther et al. 2017: 195), but they have opened opportunities for commercial exploitation, systems of surveillance and attempts at societal control. Most recently, we have seen surveillance take a participatory turn following its privatisation and commercialisation. Cohen (2014: 2) succinctly states ‘we are playing, but we are also being played’.

5Within this context of datafication, participatory surveillance capitalism and prospering rhetoric about scientific insight from big data, the practice of data shadow donation exists. Data shadow donation refers to individuals digitally gifting their collections of personal data to research or charity platforms, aka for the public good. Data shadow donation sets itself apart from traditional donation practices in that it does not refer to the gifting of something physical and finite, such as organs, money, or blood. Instead, it is a by-product of our ‘post-digital’ era of socio-technological development, whereby the distinction between the digital and the non-digital has become blurred (Berry and Dieter 2015: 2). Scholars have come to categorise the practice as a ‘pro-social behaviour’ and a form of ‘digital philanthropy’ (Taddeo 2016; Skatova and Goulding 2019). Furthermore, this relational activity engages citizens and research institutes in the value creation process, thereby challenging perceptions of data as solely a tradable commercial commodity used for ‘advertising’ (Spiekermann et al. 2015). By re-situating the citizen within the data ecosystem, the practice challenges the power imbalance that currently characterises the data-sphere (Carmi and Yates 2021). Data shadow donation thus becomes an instrument for civic engagement and, second, an expression of public value and shared control. 

6To understand the experiences of data shadow donation in more depth, an ethnographic research project of the Our Data Helps platform (Ourdatahelps.org) was conducted. Whilst the research project and findings are reported in-depth elsewhere (Sleigh 2018), this article seeks to reflect on the art installation that resulted. Taking the form of an ambient visualisation, a category of data representation that conveys information in the periphery of human attention, the work invited audiences to let their movements be followed and transformed by technology and collections of other individuals’ data. In this way, the work made visible surveillance of technological interaction. This article focuses on the content, style, and methods used to create the work and on how new technologies were experimented with and employed to creatively resolve visualisation challenges. Further, this paper contributes to discussions on data as a visuality, new media technologies, and the inclusion of the subjects and viewers as image-makers. 

Background: the research project

7The foundation of the installation discussed in this article is the explorative study of the phenomenon of data shadow donation. By focusing on Our Data Helps’ online community for data donation, the study aimed to obtain an in-depth understanding of donors’ experiences of donating data to mental health research. To give some background, Our Data Helps was founded in the U.S., and since 2016 they have been working together with data analytics company Qntfy (Qntfy.com) to accumulate personal data collections to advance social, behavioural, and biomedical research. The community shares a belief echoed in our society that data can reveal insights about the very nature of our being and behaviours (Mayer-Schönberger and Cukier 2013). As such, individuals gift the platform their publicly posted messages on social media networks (such as Facebook, Twitter, Instagram), along with wearable sensor data (i.e., Fitbit) and workout data (i.e., Runkeeper). Our Data Helps then, in collaboration with Qntfy, uses collected data to test machine learning models for mental health research. They do this to:

81. better understand mental well-being,

92. develop new ethical protocols for online media health research, and

103. support the development of preventive treatments in the field of mental health.

11Focusing on this platform, guided by the research question of ‘how is the donation of personal data to OurDataHelps.org experienced?’ and following a constructivist epistemological perspective, the research took a mixed-methods approach. First, participant observation helped gain insight into the broader themes linked to the phenomenon of data donation. Online platforms of community activity examined included the Suicide Prevention Social Media Blog (SPSMchat.com) and YouTube channel, alongside Qntfy’s social media pages and blog. Second, semi-structured interviews with OurDataHelps.org members (found during participant observation) were conducted. Study participants were individuals who publicly identified as donors, communicating this identity online in blog posts or on social media platforms. However, although they had openly identified themselves as donors, these participants still valued their anonymity. They thus requested that their interview data be presented as non-identifiable. Written informed consent was also obtained. Privacy was further protected through the use of pseudonyms.

12The study findings revealed several things pertaining to the experience of donation. For one, that donors were aware of the continuous surveillance occurring online and on social media. Regarding the question of privacy, their primary fear was surrounding how their data could be used. A key finding was also that survivor bereavement underlay the experiences of donation to Our Data Helps and that the activity of donating helped individuals feel a sense of purpose. First, as they were helping to promote suicide prevention health research; second, it made them feel their data and online activity were valuable. This finding aligns with research carried out by Skatova and Goulding (2019), which indicates that social duty and understanding that a donation has purpose underline intentions to donate, but self-serving motives have a negative effect. Moreover, although participants felt their data was valuable and partly reflected their identity, they had no idea what their data looked like, what it comprised, or even how to access or download it.

Data as a Visuality

13The research finding that participants could not interact with, feel, or explore their collections of personal data underscored the reasoning to develop a visual installation. In one of the first semi-structured interviews, amidst a discussion about the varied types of personal data that one's data shadows consist of, a participant asked: ‘And what would my data shadow look like?’ However, answering this question is no easy task, as data collections do not take a traditional physical or visual form. As ever-growing collections of personal information and meta-data, data shadows exist as post-digital artefacts at the intersection of the digital and non-digital. Often represented by 0's and 1's, data shadows get characterised by multiplicity, diversity, flux, and dispersion across various online spaces. As absent presences that transform our experiences, data shadows reflect Don Ihde's (2002) concept of background technologies. Further, as they require coding skills to identify and track, one cannot take a camera or apply traditional methods of visual ethnography. A new technological approach is needed. 

14Over the last years, we have seen data visualisation emerge as one of the primary ways in which people can access and make sense of the sheer volume of data and of the overwhelming fog of possible correlations that data engender (Crawford 2014; Kennedy et al. 2016). This popularisation of data visualisation also ensues from technological advancements (such as digital sensors, data storage, computer graphics), alongside many disciplines' advancingly data-driven and computational nature (Bryant, Katz, and Lazowska 2008). With applicability across industries, we now see myriads of visualisation tools and technologies materialise, promising to help users explore, digest, and make sense of the information and knowledge embedded within large volumes of data (Chan 2017; Raghav et al. 2016). The resulting installation’s use of data visualisation technologies illustrates how algorithms, devices, and technologies play a significant role in filtering what we see, in making new visibilities and invisibilities (Amoore and Piotukh 2015). At the same time, the work points to how computation and surveillance increasingly get incorporated into our lived environment (Manovich 2009).

15Much like the field of visual anthropology itself, at the core of data visualisation, as a descriptive and aesthetic endeavour, exists a dichotomy of aesthetics and information, art, and science, subjective and objective (Brutti 2008). We see this in data visualisation scholarship and literature, which frequently emphasises beauty for comprehension and communication (McCosker and Wilken 2014). For example, American statistician Edward Tufte (2006) equates beauty with clarity, teaching in his books such as Beautiful Evidence how attention to the graphical qualities of line, colour and form helps convey information in visualisation graphics more precisely and efficiently. Notably, this attention to aesthetics highlights a shift within the sciences, which traditionally (in the West) focused on the rational, conceptual model of knowledge, whilst aesthetic cognition (itself layers of emotion, cultural values, sensory experience, and the significance of the experience) was downplayed (Dorter 1990). Scientific anthropology has also shifted from being a monomedia academic discipline to favour subjective and reflective approaches that welcome new technologies (Pink 2006).

16The installation aimed to embody this ‘information and form’ dichotomy and represent what Manovich (2009) calls the ‘info-aesthetics paradigm’, a theoretical concept to understand contemporary culture and its relationship with digital interfaces. On the one hand, the intention was to create something that drew upon aesthetic styles and principles manifested in current data visualisation practices (such as minimalism, emphasis on interaction and form). This acknowledges that the use of computers for design and production gives rise to new trends and forms of information processing that we interact with daily. On the other hand, the technologies used (the computer, the software MaxMSP and the projector) equally determined the style of the work and the form of interaction that resulted. For instance, the projector necessitated a large flat blank wall and specific light conditions, whilst the Kinect set spatial boundaries for the interaction. Information processing thus acted as both a force of stylistic influence and a method of design. 

Exploring New Media Technologies

17As a discipline rooted in both art and science, for decades, anthropological projects and researchers have methodologically experimented with new and innovative forms of visual tools and media for data collection and presentation (Grasseni and Walter 2014). Howard (1988: 314) explains that creative practices within ethnography and the anthropological community have burgeoned from the need to cope with the complexity and richness of the anthropological situations represented. Going beyond explorations of documentary film's heuristic and creative powers, new media works have questioned and reflected on the potentials of digital, multi-sensory, linear, and multimedia formats (Remillet and Wanono 2014). This installation project seeks to contribute to this epistemic space by exemplifying and reflecting on the digital visual engagement enabled through new visual media tools. 

18As previously mentioned, the mixed-technological approach used to create the installation sought to explore the affordance of tools increasingly used for data visualisation. Specifically, the technologies used included a laptop computer, the motion-sensing device of a Kinect, a projector, and the software application MaxMSP. The latter, MaxMSP, is a visual programming language that, for over twenty years, has been used by artists, researchers, composers, software designers and more for performances and installations (Sheffield 2018). To advance my skills in using MaxMSP, I participated in creative coding workshops in Berlin. This practice of partnership and learning from coding experts reflected the collaborative nature required in the methodological toolkit of a visual anthropologist (Pink 2006). This coding and learning process further revealed how the close relationship between data and visuality positions the coder as pivotal to the knowledge generation process (McCosker and Wilken 2014: 156). As well, by experimenting with a new-media technology, I wanted to credit the expanding range of visual techniques (beyond the word cloud, box plot, network diagrams) available for presenting data visually (Pauwels and Mannay 2019). How the technologies finally worked together is documented in the below image of the Technical Rider used in the exhibition process to install the work. 

19Concurrent to developing new skills in coding was the ‘production’, or rather, the collection of materials. Informed by recent visual anthropological studies on self-tracking and its embodied sensory nature (Pink 2011; Lupton et al. 2018), participants were asked to collect, filter and share data; this formed the visual material basis of the installation. Participants self-filtered the personal data they shared, mediating privacy and confidentiality issues by promoting contextual integrity (Nissenbaum 2004). The process of filtering through data further prompted reflection and discussion about digital surveillance and what information is collected online. In talking to participants about this process, they reflected on the mass and diversity of collected personal data. One participant's data collection was over 64GB, consisting of private messages, document files, audio messages or stickers, contacts, links they had visited, social media posts they had made, all the events attended and more. This embodies what Lupton et al. (2018) terms lively data in that a) the data is about human life itself, b) the data is the basis of others' livelihoods (social-media economy), and c) data has an influence on peoples' lives. As well, it demonstrates our entanglement with technology. 

20Following the data collection, an interactive installation was devised to demonstrate the entanglement between our bodies and technologies, whilst also making surveillance visible. This followed on from the research finding that data donors experienced feeling surveilled by technologies and were conscious of the data constantly collected about them. To achieve these goals, the work invited users to immerse themselves in the experience of being surveilled by technology and other viewers in the space. However, although the work recreates the experience of being tracked when engaging with technology, it leaves the viewers to decide how they feel about it. Do they find the tracking and projected data shadow interesting or playful? Perhaps they see themselves in a new way, or perhaps it is disconcerting. Such openness to interpretation follows how the interviewed donors had different perspectives on privacy and surveillance. 

Viewers as image-makers

21Another central aspect is the installation's play upon the relationships between the viewer, the researcher, the research participants, and the technologies used. In visual anthropology, encounters between the researcher and the researched can be first traced to the 'participant camera’/‘shared cine-anthropology' of Jean Rouch and the 'participatory cinema' of David MacDougall (Valentinčič Furlan 2018). The postmodern era then prompted deeper reflection on authority, voices, and authorship (Ruby 1991). We have since seen participatory visual research methods emerge aiming to ‘recognise the power of research relationships and methods that afford some voices and not others’ (Luttrell and Chalfen 2010: 200)

22One postmodern idea of relevance to this installation is Barthes's conception of 'death of the author'. This concept highlights how although an author writes words, the readers and their interpretation construct meaning (Newton 1997). In this way, the role of ‘meaning producer’ shifts from the author to the reader. The installation reflects this by positioning the viewers in partnership with technology as the generators of the visual. The viewers’ presence and movements activate the work, and their body determines the form displayed by the projector. Video_1 demonstrates this process. Without the technology and the viewers, no form nor meaning exists. The foregrounding of user interaction highlights the importance of the viewers’ body and movement in relation to technology. Although this participation does not resolve the power relations between author and participant, it de-emphasises the author's responsibility as an image-maker.

23Furthermore, the installation's focus on users discovering and making meaning through subtle interaction, came from the intention to create an interactive public ambient visualisation. This category of data representation conveys information in the periphery of human attention, often supporting transitions between implicit and explicit interactions (Vogel and Balakrishnan 2004). Mankoff et al. (2003) describe this type of display as having the goal of not distracting or over-burdening the viewer. The work devised did this by seeking to be calm and non-obtrusive (it had no bright colours or loud sounds, and the colour of the projected form was only faintly visible when displayed during the daytime) and only revealing information for interested users. Many audience members thus walked past the work without noticing the installation at all. When no users were before the Kinect, no projection occurred, and the work seemingly disappeared. This choice for ambient visualisation also aligns with the study finding that data shadow donation was an experience that one gives attention to at first, but then fades from consciousness over time. As one participant explained: ‘I think I donated when they first started the platform, and then honestly, I forgot that I did that until you contacted me’.

24In the end, the developed work was exhibited several times; at the 2017 Conference of the German Anthropological Association (GAA) in Berlin, Germany; at the 2018 World Film Festival in Tartu, Estonia; at the 2018 International Conference of Data Protection and Privacy Commissioners in Brussels, Belgium; at the 2018 ETH Digital Health Conference in Zurich, Switzerland; and at the 2018 HER She Loves Saint Lorenzo Festival in Milan, Italy. Many artists, visual anthropologists, social scientists, and the general public and groups of students attended these events, and so too became the makers of the artwork. These various display settings and types of audiences demonstrate, on the one hand, how data visualisation makes information accessible to non-expert audiences, and on the other, how visuality helps bring anthropological projects to wider audiences (Pink 2015).

25

Conclusion

26To conclude, the installation presented in this paper sought to represent data shadows (a phenomenon set between utopian and dystopian rhetoric) and make the interconnected relationship between data, bodies, and our notion of ourselves visible. Emphasising users' interaction within the installation space highlighted how surveillance and tracking are elemental to the phenomenon. At the same time, this interaction illustrates a way for participants to co-create the visual, their engagement becoming intertwined with technology in the act of representation. Further, in exploring new media tools, the work harnessed accessible technologies, namely, the computer, projector and MaxMSP software. In doing so, it recognised how data depend on technologies and visualisation methods to be made accessible and understandable.