1 Introduction

Through the digital data traces we leave every day, we have become ‘transparent subjects’. If ‘we are data’, as the title of a recent book suggests (Cheney-Lippold, 2017), then algorithms determine our identity, our friendships and our access to news, information and ads; algorithms assign us to race, gender, and class categories, and regulate which forms of citizenship and politics we can enact. In this view, the central political effect of data is to render us powerless in the face of the corporations that extract our behavioural data in pursuit of profit and/or the governments which profile and police us at ever more granular levels by using automated data tools (Couldry & Mejias, 2019; Zuboff, 2019). At the same time, the ubiquity of digital data enables other political effects, such as their use to enhance equality or the centring of data justice and data activism in collective efforts for change.Footnote 1

Rooted in a post-foundationalist and generative orientation to (critical) discourse analysis, this chapter explores the political subjectivities which are—or can be—associated with educational activist projects on data in public life. It does so by first, in Sect. 2, reflecting on cultural and political theories of political subjectivity/subjectivation which offer important insights into digital citizenship and data activism in today’s datafied societies. We then turn in Sects. 3, 4, and 5 to a ‘worked example’, analysing the discourse of non-profit data analytics projects, run by School of Data Germany (Datenschule), which teach data literacy for anti-discriminatory practices and other socially progressive purposes. We tease out the priorities invoked in the School of Data’s website, interviews with a member of the project’s team, and selected publications on open data, data activism and anti-discrimination work. To conclude, Sect. 6 returns to the theories of political subjectivities presented in Sect. 2, and reflects on the broader implications of these understandings for data justice work.

Overall, the chapter seeks to identify which conceptualisations of ‘becoming a political subject’ are at work here. While on one level, the materials issued by School of Data Germany explicitly position this process as skills acquisition (data literacy), a broader, less controllable understanding of political subjectivation as relational self-transformation also emerges. We thus conclude by suggesting that although the contemporary discourse of ‘data literacy’ often foregrounds limited, individual ‘skills’, set apart from their context, the ways in which specific projects enact data literacy demonstrate more far-reaching, power-related, collective, organisational, relational and distributed data practices. Our analysis suggests the need to address these data practices more explicitly when theorising data literacy and/or when developing educational data projects.

2 Political Subjectivities, Digital Citizenship and Data Activism

Current scholarship on data activism assumes, we suggest, various figurations of the ‘(digital) citizen’. Drawing on conceptualisations of political subjectivity in recent cultural and political theory, this section teases out three approaches to digital citizenship.

2.1 Individual Political Subjectivity: ‘I am a citizen, I am a political subject!’

Classic liberal conceptions of political subjectivity in (late) modernity posit the citizen as a sovereign, rational, autonomous and self-determined individual (Mack & Gaus, 2004, p. 116; Rawls, 1996, p. 306). One or more nation states confer ‘citizenship’ on individuals, endowing them with civil, political and social rights. This ‘thin’ conception of citizenship, as Saward (2006, p. 403) puts it in his entry in The Oxford Handbook of Political Theory, circulates around one ‘key specific space’ (Saward, 2006, p. 403), the polling booth, which largely discounts activism as part of citizenship. Approaches including deliberative democratic theory have challenged this conception, extending it to forums of “talk, dialogue, reasoning together, becoming informed together, and making decisions” (Saward, 2006, p. 404). This view reads citizenship as not simply ‘given’ (at birth, or later through formal legal applications), but also ‘enacted’, and resonates in an understanding of ‘the virtue of active citizen participation in community affairs’ (Saward), and in the idea of democratic discourse through identifiable, ”public contestation” (Schaal & Heidenreich, 2006, p. 122). This enacting is key to understanding political subjectivity in democracies, as citizens exhibit political agency and seek to gain an enlightened understanding of political issues, and act as individuals to exercise control over their polity through political representation (Isin & Ruppert, 2015).

Current discussions of digital citizenship frequently adopt a similar approach to political subjectivity, with ‘digital citizens’ seen as sovereign, relatively autonomous individuals who achieve citizenship through their practices, including “using technology to make [their] community better; [e]ngaging respectfully online with people who have different beliefs […] [u]sing technology to make [their] voice heard […] and to shape public policy”, and ”[d]etermining the validity of online sources of information” (ISTE, 2019). This digital citizen has the ‘skills’ to ”find, evaluate, and share information responsibly, engage in constructive conversation with others from diverse backgrounds, and […] ensure their online participation is safe, ethical, and legal” (Gleason & von Gillern, 2018, p. 200). In this context, “[d]igital citizenship can be described as the norms of appropriate, responsible behavior with regard to technology use” (Ribble & Bailey, 2007, p. 10). This perspective conceives of political subjects as rational individuals who enact their digital citizenship, and thus their political agency, by behaving responsibly, inclusively and ethically in digital spaces.

Similarly, some recent approaches to data activism emphasise the possibilities for individual political agency that emerge through data assemblages; they include the promotion of a critical consciousness, grassroots data literacy and a critical imagination for creating alternative ways of engaging with data that fall outwith their elite or capitalist exploitation (Gutiérrez & Milan, 2017). This work, often drawing on phenomenology, aims to engage with individuals’ perceptions, experiences and reflections of the datafied world (Couldry & Powell, 2014; Kennedy, 2018; Sander, 2020). While vitally important for identifying how datafication is not only enforced top-down on technology users, but is made and unmade in everyday data-based interactions, this research also suggests—sometimes explicitly, sometimes implicitly—an understanding of the political subject as a relatively autonomous, reflective, deliberative, intentional individual, operating within representative democracy. Although this approach reflects on communities, alliances and collaboration as part of data activism, its analytical interest lies in the sense-making, critical orientation and actions of the individual living with data.

2.2 Collective Political Subjectivity: ‘We are here, we are loud, we take up space as a political subject!’

Alongside this subject-centred emphasis , an alternative understanding of data activism emerges in other approaches that explicitly conceptualise the political subject as a contingent, relational performativity that comes into being through the collective presence of voices and bodies in public spaces. In illuminating this conception, we draw on post-foundational writing that sees the subject as emerging from (constituted, but not limited by) situated, embodied, normative sociality.

Two key thinkers in this regard are Jacques Rancière (2001, 2013) and Judith Butler (1993, 2015). Their approaches offer us two central ideas for understanding data activism today. The first of these is the ‘distribution of the sensible’, which describes the “system of self-evident facts of sense perception” that enables something to be seen and heard, thought or said, made or transformed (Rancière, 2013, p. 7). The distribution of the sensible thus enables or forecloses perception, communication and action, consequently shaping the potential of (data) activism as framed in democracy: “Politics revolves around what is seen and what can be said about it, around who has the ability to see and the talent to speak, around the properties of spaces and the possibilities of time” (Rancière, 2013, p. 8). In this view, the political participation offered to the demos within representative democracy appears contained, domesticated, tamed; its primary function is to include and exclude, to ‘police’, in the sense of keeping people in their place (Rancière, 1999, p. 28f.). A core objective of data activism would thus be to ‘redistribute’ the sensible, i.e. to change what and who is visible or invisible, audible or inaudible, and hence who has a ‘part’ in politics (Rancière, 1999, p. 9).

A second central idea is the role of bodies in activism, and how data activism relates not only to voice but also to physical presence. Butler refers to “Bodies in Alliance”, doing street politics ‘between’ themselves (Butler, 2015, p. 77). In this understanding, activist practices primarily operate on the level of the “performativity of the body” (Butler, 2015, p. 83). Contesting suggestions that protest has migrated online, Butler writes: “Although some may wager that the exercise of rights now takes place quite at the expense of bodies on the street, that twitter [sic] and other virtual technologies have led to a disembodiment of the public sphere, I disagree” (2011, n.p.).

Interestingly, there is a slight shift in Butler’s take on mediality between this 2011 lecture, ‘Bodies in Alliance and the Politics of the Street’, and the publication of the book which grew from the lecture. In the book, the sentence quoted above on ‘twitter and other virtual technologies’ ends with “I would disagree in part” (Butler, 2015, p. 94). The claim that bodies anchor political events remains, but this shift, as well as other passages, are indicative of the fundamental entanglement between bodies in physical presence and bodies in media. In this reading, then, the political event comes into being, not only between the bodies, but even more so between the bodies in alliance and their mediation (Butler, 2015, p. 92, 94).

Again, the implication for data activism is a focus on increasing the visibility and audibility of those bodies and voices usually excluded from the political (see, for example, Daly et al., 2019; Dencik et al., 2019; Milan & Treré, 2019). When, for instance, the ‘digital citizen’ enacts itself online by witnessing, hacking, and commoning, it uses the ‘performative force’ of legal, written speech acts, such as human rights declarations or national constitutions, to ‘become’ a citizen (Isin & Ruppert, 2015, p. 122f.). The ‘figure of a citizen as yet to come’ thus has a framework for the right to claim rights (Isin and Ruppert, p. 123f.). Or when, for example, lower-income families, traditionally excluded from being seen or heard as political actors, use data to raise their voices about inequality, or to show the extent to which schools discipline African American students more harshly than white students, then data undergirds the performative action of raising allied voices and making bodies present in public spaces (Macgilchrist, 2019, p. 83).

2.3 Obfuscational Political Subjectivity: ‘I am invisible and anonymous, but I still have political agency!’

One way of radically rethinking subjectivity in the context of the digital is by questioning the apparently self-evident link between political subjectivity and the process of being seen or heard. One line of argument, for instance, defines the emergence of a political subject today through enacting a “right to opacity” (Birchall, 2016) and learning to “obfuscate” one’s data traces (Brunton & Nissenbaum, 2015). Anonymity, in this sense, is not a negative practice of hiding or protecting oneself from censure; it is a positive political practice of acting anonymously with the express purpose of creating new forms of political subjectivity. This type of political subject destabilises traditional conceptions of political participation and democracy, in which the subject performs its visibility, raises its voice, and engages in open, transparent public debate (de Lagasnerie, 2017, p. 57f.).

Some anonymous online practices, such as distributed denial-of-service (DDoS) attacks, translate the physical sit-in protest into the online sphere. WikiLeaks and Anonymous are examples of how this political subject, acting anonymously, interrupts the dominant political order, which rests on the assumption that democracy is transparent, fair and participatory, yet which reveals itself as deeply implicated in reproducing socio-economic inequalities between those who participate in decision-making and those who do not. Anonymity, in this sense, is a proactive political practice of creating invisibility and, in so doing, of making the invisible visible. Invisibility is not ‘hiding’, it is acting purposefully. De Lagasnerie (2017) observes of WikiLeaks and Anonymous:

By disconnecting the question of politics from the question of the public sphere, anonymity gives rise to a scene on which what one might call nonrelational politics occurs: politics that is affirmative and radically emancipated from all ethical considerations—in other words, perhaps, pure politics. (de Lagasnerie, 2017, p. 75)

De Lagasnerie’s argument is directed against observers who claim that anonymity is a cloak for criminality, rather than a legitimate form of political participation. In his view, anonymity democratises politics by enabling contributions to the flow of knowledge which fall outwith the prohibitions and censorship that limit the enacting of democracy in democratic states today (de Lagasnerie, 2017, p. 65). It enables subjects to speak from the inside and the outside at the same time (as do whistleblowers) and reduces the personal cost of political participation, such as prosecution and job loss. In this way, WikiLeaks, for instance, enables a ‘split subjectification’ by providing the sociotechnical tools to simultaneously act within and against an institution (de Lagasnerie, 2017, p. 103; emphasis in original). Anonymity is, in this sense, a technique for casting off straightforward subjectivation (désasujettissement or ‘desubjugation’ in de Lagasnerie, 2017, p. 72), for disidentifying and disentangling oneself from the institution in which one works, for ‘escaping citizenship’ (de Lagasnerie, 2017, p. 99) as illustrated by ‘Citizen Four’, Edward Snowden.

Active creation of anonymity and invisibility takes place at the grassroots level through, for instance, various encryption mechanisms. These processes have been described as ‘reactive data activism’, i.e. an activism that reacts to the threats posed by data surveillance and mass capture of data (Milan & van der Velden, 2016). Even if de Lagasnerie does not refer to data activist practices, these may also count for him as “practices that are freer and more selective—more and more emancipated from the psychic hold of external and arbitrary constraints” (2017, p. 72). This notion of a freer practice, however, brings de Lagasnerie back to an understanding of political (‘nonrelational’) subjectivity which is strongly intentional and relatively autonomous. His account of anonymity and obfuscation in politics, while aiming to describe a novel political subjectivity, therefore returns us to a notion of individualised political subjects. This said, it does raise the question of the workings of digital data in other, perhaps more collective and contingent, forms of political subjectivity.

3 Worked Example: School of Data

To consider these theories of political subjectivity in concrete settings, we turn now to an extended example of a data activist project in Germany. After presenting our focal case in this section, we draw in the next two sections on a selective and intensive discourse analytical reading of website materials, two interviews and further publications. In doing so, we focus explicitly on how the texts account for their political and educational goals and how they relate to political subjectivities (see Kessl, 2005).

‘School of Data’ Germany (Datenschule)Footnote 2 was initiated in 2015 as one of several specific initiatives focusing on ‘open data’ run by the German chapter of the Open Knowledge Foundation (OKFde).Footnote 3 OKF is a non-profit organisation originally founded in 2004 in the UK, which champions the use of open data as material for empowering practices. The concept of open data, encompassing the subset ‘open government data’ (OGD), has gained substantial attention since the end of the 2000s and is closely related to the free software movement and other open-source and open access initiatives such as open culture and open educational resources (OER). Hintz et al. (2018, p. 132f.; see also Gutiérrez, 2018, p. 49ff.) classify open data campaigns as a form of ‘proactive’ data activism. The OKFde website outlines the principal characteristics of open data, in line with a generally accepted brief definition, as follows:

  • Availability and access: Data should be available as a whole, at a cost no higher than the cost of reproduction, preferably as a free download on the Internet. The work should also be available in an appropriate and modifiable form.

  • Use and re-use: The data must be made available under conditions that allow use, re-use and association with other data sets. The data must be machine-readable.

  • Universal participation: Everyone must be able to use, re-use and re-distribute the data. There must be no discrimination against any persons or groups. The subsequent use may not be limited to individual areas (e.g. only for educational purposes), nor may certain types of use (e.g. for commercial purposes) be excluded. (Open Knowledge Foundation Germany, 2019)Footnote 4

To a large extent, open data stems from public sources, as governments and administrations collect and work with information that is free from ownership and is considered highly relevant to the public. Key reasons given in support of opening up data are (1) transparency, (2) social and commercial value and (3) societal participation and engagement. The OKFde website explains:

In a well-functioning, democratic society, citizens need to know what their government is doing. […] Much of the time citizens are only able to engage with their own government sporadically […]. By opening up data, citizens can stay better informed and be more directly involved in decision-making. This is […] about making a full ‘read/write’ society , in which citizens […] are able to contribute […].Footnote 5

Authorities’ objections to opening up data include the risk of making official secrets or personal/identifiable data widely available. Any disclosure of data, as in, for example, responses to Freedom of Information Act requests, will need to proceed after consideration of these concerns (Charalabidis et al., 2018, p. 1ff.). Scholars have criticised open data for opening the door to new public management methods in governance and to commercial interests that monetise open data (Dander, 2014, p. 45ff.), for depoliticising the public be turning them (us) into auditors (Birchall, 2016), or for operating as governing technologies by, for instance, not only publishing data but producing data through their ‘aesthetic practices’ (Ratner & Ruppert, 2019).

Other research about open data suggests that “this type of transparency has the potential to support the agency of datafied publics” and that it is taken as a means through which “the people should be considered again as the sovereign”’ (Baack, 2015, p. 4f.; Baack, 2018, p. 169f.). However, even the members of the OKFde team have described this ideal of direct involvement between citizens and governments as overly simplistic:

Even though the idea behind the democratization of information is to potentially allow everybody to interpret raw data, activists are well aware that the average citizen does not have the time and expert knowledge to do so. They recognize that their vision of empowerment through open data can only be realized with intermediaries that make raw data accessible to the public. (Baack, 2015, p. 6)

Given this lack of ‘expert knowledge’ on working with data, many have noted the need for educational projects in this area (cf. Wylie et al., 2019, p. 177f.). School of Data responds to this need by conducting workshops on data analytics skills for NGOs, young people and youth workers and by bringing data experts together with political activists to work on joint projects. At the time of our data generation, a team of five, consisting of one project manager/developer, one communications/press officer, two IT experts and one facilitator/coach, ran School of Data; as of June 2020, the website lists four permanent team membersFootnote 6). We conducted interviewsFootnote 7 (see below) with the facilitator/coach who had conceptualised and predominantly run the workshops. He has a background in European Studies and Public Policy, but also some programming skills (I01:7:27–31). Although the team’s roles seem to be clearly defined, all members work on multiple tasks, are politically involved and work closely with one another.Footnote 8

The team ran and runs several projects, including ‘Every School’ (Jede Schule), an online database with open data on schools in Germany, ‘Democracy Labs’ (Demokratielabore), a joint project for media and civic education for young people and youth workers, and a collaboration with the environmental organisation ‘Robin Wood’ that focuses on environmental data.Footnote 9 Of particular relevance to this chapter is a joint project with ufuq.de, in which the two organisations co-developed a report on ‘Using open data in anti-discrimination work’ (see Sect. 5).

School of Data does not cooperate with businesses or other commercial actors; its focus is explicitly on non-profit organisations. The website notes: “We at School of Data Germany want to help realise data-driven projects: We cooperate with non-profit organisations based in Germany that campaign for positive social change.”Footnote 10 Neither the organisation’s mission statement nor any other part of its website specify a definition of ‘positive’ in this context; our interviewee supplied a partial clarification by stating that cooperation with right-wing actors is not an option (I01:12:20–23). Further, irrespective of political leanings, School of Data does not accept requests to cooperate with any of the foundations formally aligned with political parties in Germany (be they conservative, socialist, green, etc.). Overall, then, the organisation’s emphasis is on independent work with civil society actors on the progressive spectrum of political activity, within which context School of Data conducts numerous workshops each year (I01:18:17).

4 Modelling Data Literacy: From Skills to Empowerment

In the early years of open data, demands and discussions relating to technical, legal and political aspects of the issue dominated the discourse (Dander, 2014). We now note a closer focus on the specific educational and social practices of activists and various groups in civil society which seek to make open data actionable for counter-hegemonic positions. A long-term goal in this context is to make data experts/intermediaries unnecessary. It is in this context that we can read School of Data’s principal societal and educational objectives, as listed on its website as ‘knowledge transfer’, ‘social change’ and ‘community building’. The website contextualises the project’s activities as follows:

[N]on-profit organisations […] often lack the required skills to use and turn data into valuable information. […] In our workshops we teach the skills needed to process data and use related technologies in an independent and self-determined way. (Datenschule, 2021; emphasis added)

A similar description appears on the international School of Data website:

We are a network of individuals and organizations working on empowering civil society organizations, journalists and citizens with skills they need to use data effectively […]. We are School of Data and we believe that evidence is power . (School of Data, 2021; emphasis in original)

The descriptions refer to ‘skills’ as something people can ‘lack’ and that can be ‘taught’ in workshops, and postulate that one set of actors can be ‘empowered’ by another. With better skills, the texts suggest, people will be able to process data and use technologies independently and effectively. Alongside this understanding of skills as something that can be adopted by individuals in decontextualised settings and then implemented in participants’ own projects, and somewhat contradictory to it, sits an alternative focus on ‘organisations’, community-building, campaigning and power. This foregrounds collaborative or structured work in groups or teams, within organisational and hierarchical societal structures.

Conceptually, School of Data uses two basic models developed ‘on the job’ in cooperation with academic researchers: the ‘data pipeline’ as a process model for data projects (I01:13:27ff.) and ‘data literacy maturity’ as a descriptive model for data literacy skills within organisations.Footnote 11 The data pipeline features prominently within the booklet Jugendarbeit im digitalen Wandel (‘Youth work in the digital transformation’; Hahn et al., 2017, p. 39ff.). The model, considered a ‘work in progress’ being developed by the international School of Data community, currently consists of seven steps for literate engagement with (open) data, involving imperatives to define, find, get, verify, clean, analyse, and present data (see Fig. 3.1).Footnote 12 In expanding an earlier version of the data pipeline, Helena Sternkopf (2017) developed a data literacy maturity model. To describe ‘individual levels’ of data literacy maturity, this model adds ‘assess and interpret’ as a further category. For the purpose of evaluating organisations, the data literacy maturity model begins with two organisational aspects, ‘data culture’ and ‘data ethics and security’, which the ‘Data Literacy Maturity Grid’ records on four levels progressing from Uncertainty, Enlightenment and Certainty to Data Fluency (Sternkopf, 2017, p. 62f.).

Fig. 3.1
figure 1

The data pipeline (https://schoolofdata.org/methodology [Accessed 2020-07-22])

These models do not promote any particular political orientation; they can be utilised for a range of goals. Critique of, or reflection on, the general texture of knowledge/power configurations within ‘data’ remains secondary. This specific understanding and practice of data literacy differs from conceptions within critical data studies, where data literacy primarily consists of a critical understanding of the societal function and impact of digital (big) data (Sander, 2020, p. 4).Footnote 13

Taken by themselves, the website and these models suggest a neutral, relatively instrumental approach to developing skills, which might find use in relation to a range of questions, purposes or un/ethical endeavours. However, as mentioned above, and as the next section foregrounds, School of Data aims at specific civil society organisations, activists, and related forms of democratic empowerment (‘positive social change’).

5 Countering Discrimination with Data

School of Data’s cooperation with the non-profit organisation ufuq.de demonstrates the ways in which School of Data envision their role within data activism. Ufuq.de provides education programmes for civic education and prevention work in the fields of Islam, Islamophobia and Islamism. A brochure that resulted from a joint workshop held by the two organisations brings together the basic assumptions and technological skills that School of Data provides with the specific interests of ufuq.de and similar organisations relating to the documentation of and work on racist and other discriminatory right-wing incidents. The brochure, Using Open Data in Anti-Discrimination Work. Approaches, Experiences, Pitfalls (Puvogel et al., 2017, henceforth Using Open Data), is available open access in German and English.Footnote 14

The ideas of participation, civic engagement and (self-)empowerment through open data contained in Using Open Data echo the priorities outlined in Sect. 3 above: The brochure sketches the ‘active citizen’ as a citizen in a participatory relationship with their government, drawing information from a transparent state:

When states make closed data transparent it can become public knowledge. In this way, civil society can strengthen its public control, demand accountability from political decision-makers and ultimately promote democratic processes. There is opportunity to actively shape society and participate in public debates. Open data, thus, serves as a tool to make the relationship between citizens and state institutions more transparent and participative. (Puvogel et al., 2017, p. 4)

Further, the brochure conceptualises open data as amenable to providing (numerical or visual) empirical arguments for campaigning, public relations and strategic decision-making in NGOs:

Practically, the information can serve as a basis for argumentation, for progress reports or for strategic organisational decisions. One way that open data can help in campaign work is by communicating complex relationships in a more transparent and differentiated way. (Puvogel et al., 2017)

The examples Using Open Data contains, however, refer only in part to open government data, focusing also on (open) data generated by those people who are subject to discrimination.

One project mentioned in Using Open Data is the Berliner Register (‘Berlin Chronicles Against Racism and Right-Wing Extremism’; henceforth Berlin Chronicles), which crowdsources data for the purpose of documenting and reporting on discriminatory incidents (Puvogel et al., 2017, p. 8). The Berlin Chronicles collect ‘racist, anti-Semitic, LGBTIQ-phobic, anti-ziganist, right-wing extremist, right-wing populist and other incidents motivated by discrimination in Berlin’s districts’.Footnote 15 These incidents are reported by citizens to Berlin Chronicles via the website, by email or telephone or in person at the drop-in centres city-wide. They are also reported by partner organisations, and collated from police reports, although prosecution is not required for an incident to be included in the statistics. The data set that emerges is both quantitative (how many incidents of what type in which part of the city, etc.) and qualitative (including narrative details of specific aspects of the incidents).

Alongside the human actors, multiple artefacts contribute to the Berlin Chronicles’ activism, including flyers and information about the project, local drop-in centres, telephone lines, websites, encryption protocols for transmitting data or for PGP-encrypted emails, free software for the data analysis, visualisation and presentation in annual reports, and guidelines for police forces on categorising discriminatory incidents. The database itself functions as the essential connective between all the elements involved. The database becomes an agent in the potentially transformative processes that those people undergo who, for instance, use the database to relate their experiences of discrimination and assault to the similar experiences of others.

Since those affected by racist or other discriminatory practices are actively involved in the data collection process, this reporting is seen as an empowering practice (see also Gutiérrez, 2018, p. 143). In the foreground, these data practices appear similar to the use of hashtags such as ‘#metoo’ to share and aggregate personal experiences of sexual harassment (Dolata & Schrape, 2018, p. 49). In the background, however, the forms for reporting discriminatory experiences as used in the Berlin Chronicles provide a much more structured dataset for further data analysis without erasing the unique aspects of each experience. Although the datasets incorporate reports by the police, the definitions and categories in use exceed those provided by the authorities. In this way, the simultaneous uncovering of individual and structural aspects of discrimination illuminates the specific role of data as a medium in the context of anti-discrimination work as highlighted by the Using Open Data brochure:

The central roles of anti-discrimination work are: to make discrimination visible and to create a public sphere. This involves presenting individual cases of discrimination or assault anonymously, but, in particular, revealing the structural discrimination behind the individual experience. […] Data makes it easier to visualise discrimination and its structural background. (Puvogel et al., 2017, p. 7)

Data are thus presented in the brochure as transcending purely quantitative operations and allowing for both quantitative and qualitative analysis. In this process, data literacy skills are helpful in refining and analysing data, recognising and visualising patterns of discrimination within datasets and, finally, communicating findings in a clear and plausible fashion (see Fig. 3.1 above). In this view, the criticisms of crowdsourced data as incomplete and not valid or representative do not necessarily identify a disadvantage of this form of data:

[…] raising awareness and empowering those affected […] often also includes the conscious decision to document subjective depictions that are not necessarily validated by other sources. In this way, reporting discrimination in itself becomes an act of self-empowerment by exposing the experience and making it visible. (Puvogel et al., 2017, p. 7)

This approach, described in the School of Data/ufuq.de materials, and put into action in projects such as the Berlin Chronicles, locates self-empowerment in the process of finding a data-based way of giving visibility to the experience of discrimination as part of structural discrimination. The empowerment of others via the reception, collation and anonymisation of crowdsourced data emerges here as a further important aspect of this type of data-based anti-discrimination work:

The collection and documentation of discrimination can often be combined with practical advice and support for those affected. […] the [local drop-in] centres can also identify support structures and suggest individual ways of dealing with these issues. (Puvogel et al., 2017, p. 7)

This observation demonstrates the link between practices of data collection and documentation and social relationships revolving around communication, advice and support. The details of the data-based anti-discrimination work described in Using Open Data highlight the deeply collective, relational practices of doing data activism.

6 Political Subjectivities, Data Activism and Data Literacy

This final section brings together this worked example with the cultural and political theory presented in Sect. 2. How, specifically, do these data practices enact political subjectivity, and what broader implications do the theory and worked example have for data activism and data literacy?

In our analysis, we have identified all three forms of political subjectivity discussed in the theoretical literature. In the interviews, mission statements, websites, publications and data literacy models, School of Data enacts the idea of an individual (political) subject. The onus is on teaching data literacy as individual skills acquisition, it is on independent and self-determining data use, and on becoming fluent in using—but not necessarily critiquing the power of—data. When these skills are acquired, the materials suggest, individuals will be empowered to challenge today’s widespread racist and other discriminatory practices.

However, with the shift from the data pipeline to the data literacy maturity model, School of Data also shifts away from a focus on individuals and towards a focus on organisations. The educational endeavour moves from individual skills acquisition to organisational development. This itself is an ambivalent reorientation. On one hand, it can seem like an elitist move to target organisations that are already active in activist practices, rather than anyone who walks in from the street. However, the organisations targeted by School of Data are themselves working within networks of grassroots activists and volunteers. If we understand non-profit activist organisations as collective configurations or communities of political practice, the move towards supporting organisational development enacts a shift from individual to collective political subjectivities. Data literacy is no longer seen as individual and context-free but as contextualised in collective projects.

The Berlin Chronicles show that inclusion in anti-discriminatory data practices can take many forms and occur on very different ‘levels’ of data literacy. These projects interlink several practices—and a range of human and non-human actors—to create a database of racist and other discriminatory attacks on which activist campaigns can draw. The Berlin Chronicles provides people experiencing or witnessing racism and discrimination with both a narrative and ‘data points’ on their experience. They may individually write an email or visit a drop-in centre (although they are equally as likely to go with a friend), but the key to the Berlin Chronicles is that each person’s experience becomes part of a collective, distributed configuration, and their personal identity is obfuscated within this configuration. It is precisely through the anonymisation and aggregation of the large number of incidents that the data become a powerful force in anti-discrimination work that redistributes the sensible (Rancière) and makes the otherwise invisible visible.

Overall, thus, in our worked example, the explicit conceptualisations of citizenship and data literacy remain broadly within the liberal, individualistic state-citizen relationship of informing, debating, voting and individual empowerment. However, in the details of recent, specific, contextualised projects in which our focal organisations turn to action, we read an understanding of citizenship and data literacy which reach beyond this framing, and enact relational, collective and obfuscational political subjectivities. These projects potentially lead to empowerment by giving individuals anonymous access to resistant, counter-hegemonic practices and thereby entry into a collective, an assemblage of human and non-human agents, enacting distributed, networked data literacy.

The specific analysis suggests implications for data literacy. First, what would happen if data literacy models explicitly framed data literacy as a collective, relational endeavour? Would they have the same uptake in society as current models do? Would they have a different effect? Second, data literacy models or teaching would need to include a self-reflexive moment on how the data literacy model or educational approach itself subjectivates those becoming literate: as autonomous, self-determining, modernist individuals; as relational, collective beings; or as individuals empowered to obfuscate (and to turn the usually negative connotations of the word ‘obfuscate’ into a positive word for today’s world in which ‘we are data’).

The example also suggests implications for data activism, which are perhaps best framed as open questions. Voice and bodily presence doubtless remain vital to activism. This said, we wonder whether the societal transformations potentially arising with datafication are also transforming political subjectivities beyond these classic forms of activism. What happens if we adopt a relational, collective approach to activism as distributed across multiple actors? How can datafying practices be thought beyond themselves, towards a more emancipatory perspective? Within dynamic data activist practices, involving humans, artefacts, symbols and texts, we recognise not only an opportunity for counter-hegemonic participation in public spaces and for the experience of self-transformation (speaking with de Lagasnerie: désasujettissement or desubjugation), but are also aware of the potential reinstatement of overarching, dominant cultural and discursive norms and orders, such as Enlightenment conceptions of political subjectivity.