From Place-Memories to Active Citizenship: The Potential of Geotagged User-Generated Content for Memory Scholarship

The unprecedented circulation of user-generated images of public spaces on social media offers memory scholars the opportunity to reconsider the role of place as a catalyst for transmitting memory. Drawing on the study of a large data set of geotagged images (including text-based metadata) from three European public squares posted on Flickr, this chapter discusses how online practices are reconfiguring the interrelations between memory, identity, and citizenship. To do so, it explores how the relationship between personal memory and historical memory of place is articulated on geosocial platforms. In this context, we propose a set of methodological insights focusing on the role of the geotag to develop understandings around the circulation and solidification of place-memory in the digital realm.


Introduction
The unprecedented circulation of user-generated images of public spaces on social media offers memory scholars the opportunity to reconsider the role of place as a catalyst for transmitting memory (Schama 1995;Basu 2013). In particular, the communicative nature of online practices constitutes an additional asset for researchers and an element of complexity. Indeed, conversational remembering is at the core of established methodologies in the field of memory studies; in this context, the orality of the interview is expected to support an engagement with vernacular forms of memory, often in antithesis with official ones (Mihelj 2013). The idea that social media could offer a further opportunity for the exploration of noninstitutional memory is undoubtedly appealing, particularly considering the potential to work with materials that are both unsolicited (by the researchers) and embedded in contemporary, everyday practices.
Drawing on the study (Arrigoni et al. 2017) of a large data set of geotagged, 1 user-generated images and their text-based metadata from three European public squares posted on Flickr, this chapter discusses how online practices of sharing images of place on social media platforms reconfigure the interrelations between memory, identity, and citizenship. The study provided an insight into how the relationship between personal and historical memory is co-articulated on geosocial platforms. Our interest in photo-sharing platforms in relation to place-making and place-memories was dictated by the broader research context 2 in which this study was carried out. The overarching research question concerned the role of digital technologies in enabling dialogue around heritage and identity issues, with a particular focus on European heritages and the current European Union (EU) crisis. Thus, besides investigating installations and platforms specifically developed by museums to support dialogue, we wanted to explore how dialogue around heritage also takes place outside of any institutional framework, without facilitation, or serendipitously, for instance, on social media. Our study was therefore guided by the hypothesis that the analysis of geolocated user-generated content on social media can contribute new understandings of the public's experience and the circulation of cultural and historical memories associated with a place. Indeed, we looked at such historical significance as constitutive of both the heritage and the memory discourse around a place, conscious of the differences but of the parallelism as well in these notions. Consequently, our key aims were to assess whether the construction of place-memories on geosocial platforms could take shape through dialogic forms; and to understand how online practices could sustain the coexistence of both aligned and oppositional approaches to official or authorised (Smith 2006) representations of place.
The analysis of selected images (or groups of images) from the data set explores different ways in which photographs of places are embedded within different processes of identity and community building. Our findings point to the activist and 'civic oriented' nature of an interesting portion of the images, in which the cultural and historical significance of place is mobilised as a trigger for debate on current socio-political issues, to express a political stance or to enhance the symbolic currency of public events such as protests and activist campaigns. In this context, we discuss a set of methodological insights employed to explore the research potential of social media beyond the increasingly popular net-ethnographic approaches (Hine 2008;Pink 2016). In our view, the specific immersive protocols of net-ethnography, 3 and its primary focus on users' behaviour might not always provide the most fitting approach to cross-disciplinary research endeavours, for instance, when addressing large and heterogeneous data sets in a qualitative fashion.
In our study, we combined visual methods with speculative and interpretative case studies to analyse large amounts of user-generated images and their metadata by framing individual images in relation to specific themes. Furthermore, we analyse the methodological importance of the geotag to develop understandings around the circulation and solidification of place-memory in the digital realm. Finally, we reflect on how exploring the conundrum of place-memory-identity on social media image-sharing platforms leads us to recognise processes of almost instant historicisation of recent public events that superimpose themselves to, and become part of, the layered corpus of cultural meanings associated with a given place. Working with unsolicited contributions posted by users unaware of our research activity also prompted us to reflect on a set of ethical issues. A further contribution of the chapter indeed concerns some ethical considerations around the use of content from social media for research purposes and the problematic necessity to make visual research data sets available in the public realm.
The next section briefly introduces key theoretical considerations around ideas of unofficial heritage and memory discourse, followed by an extensive discussion of the rationale and design of the chosen methodological approach used in this research. We contextualise and illustrate our theoretical and methodological reflections through two case studies from the data set. The chapter concludes with our discussion of the advantages and limitations of our methodologies, including the new ethical considerations arising from this investigation.

The Civic Life of Unofficial (and Implicit) Memory Discourses
The well-established relationship between memory and photography (Barthes 1980;Lury 1998;Langford 2001;Khun 2007) has been repeatedly questioned given the wide diffusion of digital photography and photo-sharing practices. This discussion has mostly focused on the shift away from the mnemonic role of photography, towards stronger communicative and identity-building functions (Gye 2007;van Dijck 2008;Van House 2011;Keightley and Pickering 2014). Although the coexistence of these different roles is generally recognised, scholarly interventions can be positioned across a spectrum, from those more strongly emphasising continuity with traditional or pre-digital uses of photography, and those stressing the current communication-oriented and ephemeral nature of photo sharing (van Dijck 2008; Keightley and Pickering 2014). These debates however tend to concentrate on individual behaviours and the ways in which images support the remembering of a personal past. Our research, on the other hand, questions whether digitally distributed and shared photography have a broader, civic dimension beyond the sphere of its production. We ask whether a place-centred perspective, achieved by focusing on geolocative content, can usefully contribute to the idea of a distributed, layered remembering and to the assessment of its civic and social relevance?
The civic potential of shared images of place speaks directly to the contested relationship between personal and collective, independent and institutional memory. Memory and heritage narratives of the past are now recognised in their discursive, multi-actor dimension comprising official, authorised voices (Jovic 2004;Smith 2006) alongside alternative, dissonant, bottom-up perspectives. For instance, the potential of social media to challenge traditional heritage approaches is acknowledged by a growing body of literature addressing the value of user-generated content to the work of institutions of memory (Giaccardi 2012;Ridge 2014). Oftentimes, this literature tends to privilege examples where the plurality of voices gathered through digital platforms becomes integrated within the institutional framework. For example, Galani and Moschovi (2015) provide a critique of these assimilation approaches of user-generated and contributed photography in art museums. Moreover, there is an emphasis on clear-cut dichotomies separating official, dominant narratives, which find expression in a number of media, from neglected, repressed ones, belonging to the side of the defeated and the oppressed seeking recognition.
Alongside clearly oppositional official and unofficial voices, social media platforms have progressively become hosts to a variety of other voices that do not explicitly operate in opposition to official narratives. Rather, they may reflect and remediate institutional narratives as well as more independent or individual ones. What is significantly different is that these accounts are neither initiated nor facilitated by museums or other institutions nor are they integrated in their institutional archives and narratives. More regularly than not, they are not even consciously recorded as objects of memory. Rather, their relevance towards place and memory-making is the unintended output of routine practices shaped by mobile and social media (Hand 2016). In this sense, these voices are conceptualised as being serendipitous. At the same time, it is possible to suggest how some of this content mobilises and incorporates the past for the sake of current, contingent activist or civic initiatives, suggesting a transformative and actionoriented dimension for social media memory-work.
The role of memory in influencing social movements and civic action has been investigated through interdisciplinary approaches that also address the way images are used to construct narratives and inspire social performances (Doerr 2014). Addressing the realm of visuality as a public space for debate, where memory circulates as a multidimensional category of culture (Doerr 2014: 210), these approaches provide us with a viable point of departure for our analysis. Nevertheless, we encounter here the same focus on silenced, neglected memories that was identified as a crucial divide in heritage scholarship (see, for instance, Vergo 1997; Golding and Modest 2013;Winter 2013). Our data set, by contrast, incorporate and merge content in a way that prevented any clear-cut distinction between oppressed and dominant narratives. Rather, our data set demonstrated that these distinctions are not so polarised but blended within mundane, contingent attitudes to personal and collective memory. Indeed, the behaviours identified from the data set are closer to a conceptualisation of civic engagement associated with ideas of vernacular, everyday creativity and cultural citizenship (Burgess et al. 2006). These scholarly discourses re-appropriate the Habermasian notion of episodic, occasional publics (Bruns and Highfield 2016) to articulate the existence of temporary opportunities for debate and mobilisation embedded in a range of mundane sites and a variety of cultural practices. In other words, the civic engagement tendencies we identify associated with photo-sharing of place-memory objects contribute to a line of inquiry attributing value to the ordinary, the personal, and the experiential in public discourse (Hauser and McClellan 2009).

Methodology
The serendipitous and fleeting nature of the interactions we explored in the previous section also poses some significant methodological challenges. How can we study the opportunistic and fluid nature of visual expressions on social media image-sharing platforms in relation to place and place-specific narratives? What constitutes an appropriate data set? What data collection and analysis approaches have the capacity to prioritise the place-specific over the personal and what are their implications?

Designing the Data Collection Strategy
Because of our interest in the interplay between official and unofficial narratives in the mediation of public spaces through user-generated photography, our study prioritised the collection of user-generated images on freely available image-sharing platform such as Flickr. The scope of the images was shaped by both our research interests in European heritage(s), identit(y/ies), and place(s), and more particularly the acknowledgment of the role of public spaces, such as urban squares, in shaping the cultural identity of an area (Giddings et al. 2011: 203). Within this context, our study-aggregated images geotagged in three European squares: Loreto Square (Milan, Italy), Old Eldon Square (Newcastle upon Tyne, UK), and Kotzia Square (Athens, Greece). These sites are characterised by layered histories and mixed uses, including connections to antiquity, the Second World War and war memorials, markets, and commercial activities alongside administrative functions, such as the hosting of municipal buildings.
The aggregation of images was achieved through the appropriation of a commercial data mining tool, Geostream (Lamprianidis and Pfoser 2014), with the capacity to gather online images with a selected geolocation. 4 Although the Geostream aggregator tool was originally developed to gather data from several different social media platforms, only some of these were image-based-the main image-sharing platform aggregated by Geostream was Flickr. 5 The entire data set comprises 15,061 images uploaded between 2003 and 2016, divided as follows: • Old Eldon Square: 9498 images (5 from Wikimapia, the rest from Flickr) • Loreto Square: 3373 images (4 from Wikimapia, 1 from Eventful, the rest from Flickr) • Kotzia Square: 2190 images (2125 from Flickr, 30 from Wikimapia, 33 from Eventful) The images were accompanied by textual metadata, partially automatically generated (by the providers or the cameras), partially created by the users, and including title, tags, comments, geolocation, and dates (of picture taken and/or uploaded). The metadata was not integrated as a mere corollary element in our analysis. Indeed, it is customary in photo-sharing practices online to present images and texts as an inseparable aggregate. The metadata contributed to confirm or clarify the subject matter of the picture and provide cues and details to support the interpretative, and sometimes speculative, work of the researcher. By speculative, we mean that our analysis was often based on suppositions, conjectures, articulations of meanings grounded only in the visible cues, without actually interviewing the users, for instance.
Although the Geostream aggregator tool provided easy access to usergenerated images on Flickr, this choice came with a set of challenges and limitations. First, the Geostream data provides no background information on the users, thus removing a layer of contextualisation that would enable the interpretation of each image in association with the ongoing posting behaviours of its creator. Additionally, Geostream only aggregates photographs for which users have enabled both download and geolocative features. The fact that such features can either be consciously selected case by case or automatically enabled by users, complicated our capacity to speculate about the specificities of our data set when compared to other Flickr images of the same location, which were not part of the data set because users had elected to make them not downloadable. These limitations all together, however, helped us in maintaining a strong focus on place and its representation.

Analytical Approach: Accounting for Place in the Data
Scholars have identified a shift in the evolution of locative media from an initial phase aimed at encouraging agency in the physical space, to recent practices of performing presence and building personal identities (Drakopoulou 2017). We argue however that this performative process of identity building through geolocative features on social media also inevitably contributes to place-making through personal narratives even though this might not be the main aim for the users. If we assume that geotagged images contribute to the ongoing formation of personal digital archives of the recent past (Drakopoulou 2017), then a question about the specificities of the geotag as an organisational logic must follow. Indeed, our investigation was also guided by the hypothesis that retrieval possibilities anchored to specific locations could play differently than other components of the metadata. Some commentators address the organisational logic of the database to suggest caution in establishing the relationship between online photo-sharing and memory work. Pointing at the agency of the algorithm in determining the results of query-based search, Van Dijck (2011) and Schwartz (2014) emphasise how new orderings are performed every time, generating unique and ever-changing groupings of images that invalidate conceptualisations of social media as archives. As opposed to archival order and structure, social media content responds to the logic of flow, remix (Wilson 2009), or disordered 'miscellaneity' (Weinberger 2007). The use of tags, in particular, demonstrates users' awareness of the variable possibilities of distribution for their images, so that their presencing can be multiplied and extended to 'potentially unknown streams of unrelated images' (Hand 2016). In contrast, we argue that the geotag behaves differently from more ephemeral tags. The link to one place could provide more stability in the ordering or aggregating principle for photo-sharing content, one that can be especially suitable to be associated with memory practice and place-making processes.
Our qualitative investigation of the data set involved keyword-based searches; annotation and thematic categorisation of images; and, subsequently, a focused analysis of selected images considered particularly emblematic of specific behaviours or tendencies. In the first stage of the study, the visual inspection of the images was combined to keyword-led searches associated to conventional heritage and memory terminology (for instance: culture, cultural, monument, statue, memorial, heritage, memory, identity, tourism, place, public) as well as terms more specific to the heritage and historical character of each square. This process led to a set of categorisations and to the identification of recurrent themes and users' stated attitudes. More specifically, some images reflect a mobilisation of local heritage narratives and historical events to trigger debate on current socio-political affairs or to express a political stance. Other images contribute to a transformation of the cultural significance of the place, the way it is remembered, by associating it with contemporary collective events such as manifestations and activist campaigns. The following sections present some initial examples to illustrate how the research team engaged with the analysis of specific images. For this chapter, we will be focusing only on images geotagged in Loreto square, Italy.

Introducing the Case Studies
The main character of Loreto Square in Milan, today, is that of a busy traffic junction and metro interchange, mostly inaccessible to pedestrians, who can only move around its perimeter or through subway underpasses. Since the mid-twentieth century, it has marked an implicit boundary between the core of the city and its peripheries, which are now heavily multicultural. Despite the lack of remarkable signs of the past or historical buildings, the square is a relevant site for public memory. In fact, it was theatre of two infamous events during the Second World War. The massacre of 15 anti-fascist Partisans by a Fascist division in 1944 is an event now memorialised through a plaque and a sculpture in a nearby street. Moreover, on the 29th of April 1945, Loreto Square hosted the public display of the corpses of Benito Mussolini, his mistress Claretta Petacci and 18 high-ranking fascists, who were hung upside-down from the roof of an Esso petrol station. The absence of any interpretative sign or memorial to mark this event is commonly associated with a general difficulty of Italian institutions to deal with a fascist heritage (Mitterhofer 2013), which is still a catalyst of tensions and a strong reference for current political realities. While our keyword-led search demonstrated a limited use of heritage and memory vocabulary in the metadata (Arrigoni et al. 2017: 13-17), a deeper exploration of the images showed that the historical identity of the square was often implicitly addressed as a reference to comment on current socio-political affairs. A broadly recognised 'self-representational' mode of using photo-sharing platforms implies that users selectively curate the range of places and experiences they decided to associate with their profile, to present an ideal version of themselves (Malinen 2010: 381-383). In this context, the processes of identity building and identity presentation can be performed by taking a political stance or expressing a view on current issues.
Our first case study is exemplary of this behaviour, and it is introduced to illustrate how historical memory is used in everyday life beyond any heritage specific framework. Most of the empirical studies on the relationship between heritage and photo-sharing platforms address groups and communities dedicated to documenting and sharing images of sites or landmarks that are explicitly framed as significant carriers of cultural or heritage value (see for instance Garduño Freeman 2010; Terras 2011). Even if these practices can be described as bottom-up, participatory and unofficial, they still adopt conceptual frameworks, definitions and vocabulary typical of institutional and professional discourse; this issue has been addressed for instance by Taylor and Gibson (2017) in reference to the democratising barriers implicit in digital heritage initiatives. The particular filtering allowed by the geotag in our methodology, by contrast, allowed the exploration of a defined place without focusing only on thematic groups or pages already dedicated to the significance of such place. Our data set, indeed, aggregated images from dedicated pages too, but alongside occasional ones, capturing our chosen location as part of behaviour not intentionally framed as heritage or memory-related. Our first case study then demonstrated that historical references emerge in mundane, everyday processes of communication that cannot be defined as intentional bottom-up memory or heritage work, but in which heritage and memory, somehow, filter through.
The second case study complemented the first one by illustrating a different opportunity offered by our geotag-oriented methodology. Here we can see how working with this kind of data set enabled a dynamic approach to the investigation of the memory and cultural significance of a place. One common approach would have been to concentrate our effort in the identification of signs and references to a specific historical event that constituted the dominating factor in the historical significance of a place (in the case of Loreto square, obviously we refer to the events of the Second World War). Once again, working with the Geostream data set enabled a different perspective and a more dynamic and layered exploration of the place. In fact, we were able to detect the evolving nature of the historical characterisation of the square, constantly responding to new events and collective experiences.

Place-Memory and Contemporary Political Debate
6 Translated: Milan big Milan deparmentstore lombardia (the region where Milan is located) benito Mussolini rsi Lombardy upim shop theduce loreto square nokia 6070 repubblicasocialeitaliana (the neofascist party) department store clarapetacci. fiti establishes a comparison and continuity between Mussolini and Berlusconi (the controversial ex-Prime Minister and founder of the centre-right party Forza Italia). It is impossible to determine whether one of the users uploading this image was also the author of the graffiti. However, both titles express approval, sympathy, or support for its message: 'Someone Buy Him a Oneway Ticket' and 'Spontaneity of Gate'. Additionally, because the image only depicts the shutters, it is not possible to ascertain whether the graffiti is actually located in the proximity of Loreto square, or if the geotag was applied to connect the location of the garage with the location of the hanging, therefore reinforcing the link to this historical reference. The fact that there are two images depicting this image and both are geotagged in Loreto corroborates the idea that this is also the real location of the graffiti. One of the things that attracted our attention to this example was the role of the geotag in reinforcing and maintaining the connection between the symbol of the gallows and Loreto square. Without the locative data, the connection between the historical events referenced by the graffiti and its place would be lost and perhaps ignored by most viewers retrieving the image through other, variable queries ( Fig. 8.1).
These examples show how an awareness of wartime events that took place in Loreto square were not feeding into claims about the heritage value of the place itself, but rather were embedded into mundane political commentaries and processes of identity building and identity performance. Indeed, by sharing and commenting on the gallows graffiti, the users proffer a political stance by making clear to their audience their disapproval of Berlusconi. Significantly, this example demonstrated how current struggles and political divisions in Italy are contextualised in relation to a not fully digested past. The symbol of the gallows in Loreto square is mobilised as a living, generative, and highly malleable memory object. The users were not simply making a comparison between two political leaders, nor were they warning a contemporary audience to avoid the mistakes of the past (a frequent approach to mobilise historical memories in the present). Rather, they were suggesting an uninterrupted continuity with a perceived right-wing, authoritative and manipulative approach to political leadership.
italy milan church grande nokia san italia milano 1988 almirante moustache romano chiesa tribute movimento piazza secretary mass 2008 homage comrade piazzale lombardia loreto romancatholic italiano msi lombardy fedele catolico tributo nazionale messa alleanzanazionale destra sociale camerata 6070 sanfedele piazzaleloreto omaggio esempio alleanza seguire segretario nokia6070 giorgioalmirante movimentosocialeitaliano nuevadestra. 7 Such a detailed tagging, together with the mundane composition of the picture itself, led us to speculate on the dialogic intentions of the user. Indeed, the picture seemed to highlight the fact that fascism is still celebrated by a portion of the Italian population and that this is, according to the user, surprising, unsettling, or at least a matter of debate. Rather than being posted to capture and share a special moment, or an aesthetically interesting view of the city, we interpreted the reasons for uploading this picture in relation with the intention of stimulating comments, probably trying to find out if others too feel the same way about the persistence of fascism in Italy. This hypothesis is naturally corroborated when considered in connection with Loreto's square's history. Thus, the photo itself is only incidentally becoming an object of memory, as its main function, for its author, is in the present: an instigator of political debate on the current political landscape. Nevertheless, the locative attribution of the picture suggested how the historical memory of Loreto square becomes mobilised to enrich the discussion and reinforce the perception of the continuity between the fascist regime and its contemporary manifestations. Ultimately, this reflection hints at the idea that one of the implicit meanings of the picture is that despite what happened in Loreto square, fascism itself is not dead.

Activist Assemblies and the Living Memory of Place
To complement the contributions offered by the first case study, the second case study focused on more recent events but equally showed the potential to be considered of historical, or at least collective, relevance. In 2011, about 500 Egyptian residents in Milan gathered in Loreto Square 7 Translated: italy milan church grande nokia saint italy milan 1988 almirante moustache roman church tribute movement square secretary mass 2008 homage comrade square lombardia loreto romancatholic italian msi lombardy fedele catolic tribute national mass alleanzanazionale social right comrade 6070 sanfedele piazzaleloreto homage example alliance follow secretary nokia6070 giorgioalmirante movimentosocialeitaliano new-right-party.
to march in solidarity with their compatriots fighting against President Mubarak. In our exploration of the data set, we analysed pictures of this event within a broader category of images, including a variety of collective, political, and activist assemblies characterised by temporary and visually striking occupations of the public space. Some of these, such as in the case of the pro-Egypt protest, were local instantiations of transnational movements, signalling the role of social media not just in coordinating and facilitating the organisation of the initiative, but also in providing a transnational dimension to its documentation and its potential to shape public discourse. In our speculations around the reasons for sharing and geotagging pictures of protests and activist interventions, we identified the desire to extend the reach of public action to an online audience, therefore generating civic awareness and encouraging further mobilisation around the issues at stake. Indeed, social media contribute to provide a legacy to otherwise ephemeral events, simultaneously affecting the perception of place. Additionally, the presence of the geotag can suggest how users wanted to demonstrate their physical presence in the protest, and therefore reinforce their association with determined beliefs and values, enhancing the civic dimension of their public profile (Fig. 8.2).
Often these assemblies do not have any specific relationship with the history or historical significance of the location at stake, which is chosen mostly because of its central position in the city. Loreto Square in particular was chosen as the starting point for this protest because of its proximity to the Egyptian Consulate. Our study, then, allowed us to explore how places constantly gain new collective meanings, as Loreto Square and the surrounding area become remembered for a growing community as well as the site of the pro-Egypt protest. This process of constituting and layering the cultural and historical significance of place is always inevitably mediated. However, what is specific to the role of social media in providing a channel for such process is their immediacy. It is then possible to talk about a form of accelerated historicisation and heritagisation, which is nevertheless the by-product of practices that, in the users' intentions, are primarily about promoting their activist cause and presenting their own politically engaged identities to their followers.
This research evidenced how conducting memory research on photosharing platforms meant embracing the idea of memory-in-the-making, working with a constant unfolding of meanings, traces, and appropria-tions. The role of camera phones in documenting extraordinary, publicly relevant events and generating participatory networked archives has recently received significant scholarly attention (Caswell 2009;Hoskins 2011;Andén-Papadopoulos 2013). In this context, the almost complete contiguity between the action and its online 'archiving', observed in our study, is an emerging feature of current digital culture that is also becoming increasingly relevant to institutions of memory (see, for instance, the Rapid-Response collecting initiative at the V&A Museum in London (Millard 2017)). In our examples, Loreto Square becomes remembered as a lived place, with the geotag performing a key task in sustaining (and making visible) the complex ramifications of this place-making process.

Discussion
This chapter illustrates the benefits of conducting place-oriented memory research on photo-sharing platforms. As a realm where individual and autonomous records of the past intersect with those aligned with institu- tional values, and develop new, self-organised accretions, social media offers unprecedented opportunities for memory scholarship. This platform is particularly relevant in the context of a deeper interpenetration (and loss of distinctiveness) among the classic categories of individual, collective, and cultural memory (Hoskins 2009). Whereas netethnographic approaches holistically observe circumscribed communities or categories of users, our study opted for a set of methods directly responsive to the geolocative nature of our data. We suggest that our methodology complements existing approaches by focusing on aspects of memory work that might be overlooked by current ways of analysing social media content. By focusing on place (as opposed, for instance, on individuals), we were able to emphasise how the cultural and historical significance of a location is associated with uses of memory in the present, is made actionable, and becomes oriented towards issues of civic and public relevance. Thus, rather than observing processes of memorialisation or consecration of public spaces that were theatre of historical events, we place our attention on mundane but impactful everyday practices. These offer an understanding of how heritage and memory are actually mobilised outside of institutional frameworks, and contextualised within current, urgent, and contingent political debates and civic imagination.

Ethical Issues Around Public Data Sets
Working with content extrapolated from social media platforms requires us to reflect on the ethical challenges associated with interpreting personal content and making it visible beyond its expected online circulation and context. Several reasons necessitate researchers to make research data sets available in the public domain, including enabling further research by other scholars, and satisfying the requirements of funding bodies. In our case, a significant task of this research activity involved developing an interface to the data set that would allow us to provide access to the data set 8 without overstepping the restrictions of use and the specific privacy concerns of the source photo-sharing platforms. To achieve this, we created a database and a web interface that provides access not to files of the images stored locally but to links of the images in their original location in Flickr and the other providers. This database enabled us first of all to maintain an ethically acceptable approach to the way personal images are used for research and, consequently, inserted in novel circles of distribution. Additionally, it affords the data set the dynamism that characterises the source social media platforms and reflects the changes of status of aggregated content. Importantly, this meant that if a user decides to delete or disable the public settings of a certain image, such image would also disappear from our data set, acknowledging the agency of the user in the aggregated data set.

Geolocative Content and Memory Work
From a methodological point of view, the Geostream study demonstrated the fruitfulness of establishing constraints and adapting our methods to tools with specific parameters. Even though we initially embraced the geotag only as a way of gathering images of the same place, through the research process, we realised that it could also become the object of more substantial attention, and lead to further discoveries. Indeed, the idea that the geotag maintains the role of anchoring point for layering heterogeneous uses and meanings of one place across relatively extended temporalities is intriguing but may also require further investigations. As mentioned above, the circulation and remediation of memory on social media is subject to combined human-machine and database logics (van Dijck 2011; Schwartz 2014). Consequently, memories are constantly embedded in different contexts and streams of significance depending on the combination of particular queries and the algorithms. Differently from ordered archives, social media platforms often allow, and are built on, more serendipitous, fleeting and mostly unexpected encounters. These are accordingly synthesized by users through the association of image and text, in the form of a brief commentary, reflection, and consideration aimed at a broad audience of 'friends' and followers. In this extremely dynamic and fluid manifestation of memory work and memory circulation, the geotag can be seen as an anchor point and an ordering criterion that maintains certain properties of traditional archives (e.g. order and stability of the record). In fact, it aggregated the layered and multifaceted memories of a place, performing a role not dissimilar to that attributed to place itself, as contributing to the stability of material things and to the collective thinking and beliefs of the communities inhabiting it (Halbwachs 1977: 232). In other words, the geotag evidences and reinforces the potential of place to make memories traceable and condensed online. Further, it does so through networks of users with (presumably) a direct experience of the selected place.
One of the methodological advantages of focusing on geolocative content is that it affords the non-hierarchical 9 coexistence of a plurality of meanings and experiences. Through the multiple contexts within which the same photo or the same event is presented, the idea of place-memory can be investigated through composite strata. These practices, in some ways, stand in sharp contrast with one of the most established ways of connecting memory and place, the notion of the lieux de memoire (Nora 1997) which tends to emphasise one layer only, one point in the entire life cycle of a given location. By contrast, the methods suggested in this chapter enable explorations of the heritage and memory value of sites not explicitly memorialised or framed as historical ones.
A stratified geography of memory emerges here as the often unintended outcome of photo-sharing practices primarily aimed at presenting or performing political identities. Place is featured as the site of lived experience and the canvas where history is constantly actualised and reinscribed onto current struggles. A geolocative perspective encourages us to move beyond distinctions between private and public memories (a common feature of mediated memory according to recent scholarship, see Hoskins 2011: 272). In this framework, personal instances of social interaction may acquire documentary, historical, or heritage relevance. agreement No 693289). We would like to thank our colleagues, Dr. Lia Galani and Orsalia-Eleni Kassaveti (National and Kapodistrian University of Athens, Greece), and Manos Tzagkarakis (Talent Information Systems S.A., Greece) for their contribution to the realisation of this research.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/ by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.