Digital archives of memory institutions are typically concerned with the cataloguing of artefacts of artistic, historical, and cultural value. Recently, new forms of citizen participation in cultural heritage have emerged, producing a wealth of material spanning from visitors' experiential feedback on exhibitions and cultural artefacts to digitally mediated interactions like those on social media platforms. We discuss the problems of integrating citizen experiences in cultural heritage archives. There are good reasons for institutions to archive people's responses to cultural objects and to study the impact on knowledge infrastructures. By analysing the case studies of the EU-funded SPICE project, we argue that a knowledge organisation system for “data journeys” can help disentangle problems that include distribution, sense-making, ownership, sensitivity, privacy, and rights management. Cultural heritage, citizen curation, data journeys, knowledge graphs 3 “A museum is a not-for-profit, permanent institution in the service of society that researches, collects, conserves, interprets and exhibits tangible and intangible heritage. Open to the public, accessible and inclusive, museums foster diversity and sustainability. They operate and communicate ethically, professionally and with the participation of communities, offering varied experiences for education, enjoyment, reflection and knowledge sharing.” See https://icom.museum/en/resources/standards-guidelines/museum-definition/ (Accessed 26/06/2023) Workshop Proceedings
Digital archives of memory institutions are typically concerned with the cataloguing of artefacts of
artistic, historical, and cultural value. In 2015, The Warwick Commission found that “the gap in
participation between the white and BAME population is widening” and reported that UK residents
from higher socioeconomic groups accounted for 87% of museum visitors. The challenge was not only
accessibility but also “a mismatch between the public’s taste and the publicly funded cultural offer” [
Experiencing cultural heritage requires engaging with the so-called cultural background (historical,
social) but also, and possibly more importantly, relating the objects (artefacts, artworks, …) to our own
experiences and, eventually, the experiences of others, which cannot be forced into a unique, objective
meaning [
Crowdsourcing initiatives have been considered by cultural heritage archives as a way, for example,
to enrich the library metadata [
2023 Copyright for this paper by its authors. CEUR
ceur-ws.org
communities outside the mainstream, but those are developed independently from cultural heritage
institutions [
Citizen curation [
Design Museum Helsinki (DMH)5. Developed by Aalto University in collaboration with the Design
Museum Helsinki, the Pop-Up VR Museum aims to bridge the physical accessibility gap, making it
easier for people to experience art and culture [
Galleria d’Arte Moderna (GAM)6. Developed by the University of Turin in collaboration with
Fondazione Torino Musei7, GAM-Game allows visitors to create stories to document their moods and
reactions to the contents they encounter during the visit [
Hecht Museum (HM). Lead by the University of Haifa, the Hecht’s Museum8 case study focuses on
engaging with students of secondary school(s) where historical artifacts are linked to an historical event
(the Galilee rebellion); participants elaborate short essays making emerge diverse opinions regarding
historical and national issues [
Irish Museum of Modern Art (IMMA)9. The Deep Viewpoints system, developed by The Open University, is based on the slow looking methodology, where the experience of the artworks is mediated through prompts and questions, and user responses are collected and confronted, and visitors themselves propose their own prompts for other users, effectively scripting their own citizen curation 13].
Museo Nacional Ciencias Naturales (MNCN). Padaone games10 develops “serious games” for
cultural engagement. In this case study, a treasure hunt game mixes puzzles, quizzes and questions
linking objects in the natural history museum to themes of environmental sustainability [
A recent survey on the topic [15] covers requirements, state-of-the-art technologies, and infrastructures for citizen curation. It characterises citizen curation from the view of user roles and devises a general workflow11. Fig 1 illustrates a typical citizen curation scenario. It starts from a cultural heritage digital asset belonging to some original author or copyright holder, then moves to how it is collected and curated by a cultural heritage institution, and finally, how it is used in a citizen 4 http://spice.kmi.open.ac.uk 5 https://www.designmuseum.fi/ 6 https://www.gamtorino.it/ 7 https://www.fondazionetorinomusei.it/ 8 https://mushecht.haifa.ac.il/ 9 http://imma.ie 10 https://www.padaonegames.com/ 11 Here we focus on analysing how to describe those systems as data journeys and why this is useful. We refer the reader to the survey article [15] for an in-depth analysis of the technologies that could support the integration of citizen experiences in cultural heritage archives. engagement system, typically produced by a third-party organisation, for example, a company active in the tourism sector. We can identify four major roles: the owner, the custodian, the builder, and the end user. The owner is the copyright holder of the cultural heritage asset; the custodian is the intermediate organisation (a museum); the builder is the company that produces the system that processes the digital object; and, finally, the visitor is the end user.
It is interesting how citizen curation turns this model upside down. In a citizen curation application, all the above is still valid, but, in addition, the roles apply in reverse. Visitors produce initial responses to the artworks and may be acknowledged as authors (copyright owners) who delegate their content management to a platform provider. Thus, such novel, digital-born cultural heritage is handled by an intermediate organisation (e.g. the tourism company) that has to oversee the content produced and, for example, ensure its quality and compliance with regulations (e.g., the content itself does not violate any law). This organisation acts as a custodian, with a relationship to the citizen like the one that ties museums to artists. The application provider then passes the newly acquired content to the museum professionals that select and curate citizen responses -- building on the content received and archiving it in the collection management system (or its extension tailored to citizen contributions). Finally, curators, researchers, and historians are the end users of this journey following citizen contributions into cultural heritage archives [15].
It is straightforward to recognise how such a scenario raises several problems with information management that go well beyond what is typically supported by collection management systems. • Sense-making. These systems generate a wealth of information that is strongly dependent on the method used for their collection. To make sense of this information, capturing the modality of their production is fundamental, as well as explaining citizen curation applications alongside the generated content. • Ownership. Users produce original content, and there is a question of whether they should be recognised as authors and whether such recognition should be promoted by the cultural heritage institution, for example, with attribution statements. • •
Monitoring sensitive content. The produced content has the same characteristics as social media data, which can include inappropriate or sensitive information. Thus, monitoring and moderation is essential before it is included in the archive or repurposed. In addition, the content may include personally identifiable information that may violate privacy regulations. Crucially, museums could be seen as being responsible for or endorsing opinions found in user-generated content.
Terms of use. Terms and conditions may restrict the use of digital images, and users should be informed of any restriction when designing the experiences (e.g. including artwork in a codesign workshop) and when the contributed content is collected and associated with the original artwork. In addition, these systems aim to support citizens in sharing their contributions with the museum and each other, opening essential issues regarding rights and terms of use of the generated content.
Considering this research program, it is an open question of what type of knowledge representation could support citizen curation. As we rely on complex systems to support the management of cultural heritage collections and digitally mediated systems to enable innovative engagement applications, it becomes vital to equip underlying infrastructures with means for monitoring, capturing, and explaining what users do with those systems. However, what does this mean from the knowledge organisation standpoint?
Provenance is a well-established notion in museum curatorial practice, where it is related to ensuring the quality and lineage of an object as part of the acquisition management phase. This idea has been borrowed by information science research, which reformulates it as the problem of describing how a certain information object has been produced, who is responsible for it, and associated usage requirements. Digital library research stresses the importance of understanding the context in which catalogue metadata is being produced and the impact that such background has on how the catalogued items are perceived [16]. Provenance plays a key role in the web literature [17], considering the people as content creators and advocating for integrating this feature in the semantic web [18]. In this declination, provenance becomes a relevant concept for us, where the assets produced by citizen curation activities are supposed to be managed as first-class objects in museum archives. This line of research is being resurged recently in the context of data studies, with the notion of data journeys, defined it as the ``movement of data from their production site to many other sites in which they are processed, mobilised and re-purposed.'' [19]. The work in data studies emphasises the difficulty of empirically understanding data journeys because of many perspectives12. In [20], a layered semantics perspective to the definition of data journeys is introduced: “a Data Journey is a multi-layered, semantic representation of a data processing activity, linked to the digital assets involved (code, components, data)”. 2.2.
The Data Journeys Ontology (DJO) [20] capitalises on semantic web research on workflow representation and reasoning [Error! Bookmark not defined.21,22,23]. DJO identifies two abstraction layers: a data node graph linked to data objects (resources) and their changes within the process and an activity graph connecting high-level operations. Fig. 2 shows the hierarchy of classes defined by the ontology. Apart from the type Data Journey, the ontology defines two main top-level types: activities and data nodes. Activities represent operations performed on the data, while data nodes qualify roles of resources in the process, either files, variables, or temporary objects.
Thus, a multi-layered data journey allows a multiplicity of perspectives that can be overlaid to describe the process. This multiplicity can help capture the context around a data journey while still allowing for computational analysis. Layering representations allow linking them to the concrete assets involved (e.g., with URLs) but also defining incremental intermediate abstractions.
Fundamentally, here we argue that the journey a citizen curation object goes through, its lineage or provenance, is a powerful way of describing citizen curation applications. We pose two main questions: (1) how do DJO components relate to citizen curation? (2) what additional representational layers are needed to capture the complexity of citizen curation? 12 It is interesting for us to note how in the book, the one chapter touching on cultural heritage is focused on the issues of forgery and attribution in the arts. Provenance is indeed the conceptual ancestor of data journeys.
We can use the five case studies of the EU project SPICE to reflect on the applicability (and utility) of data journeys. Table 1 shows a summary of our analysis of the five SPICE case studies. Although DJO specifies both classes and relations, here we mainly focus on discussing activity types, which seems the most reasonable way to approach the problem and leave an analysis of data node relationships to future work.
Thus, we can look at how DJO activity types relate to citizen curation:
Retrieval. This activity is performed in all applications scrutinised, who need to access items in different ways; it can relate to finding an artwork within a collection or collecting user responses.
Preparation. Most scenarios require a content curation phase, where developers (builders) setup
additional metadata required by the application, which can be different from what is typically
considered by cultural heritage archive metadata (it is the case of VR objects that are newly acquired).
In some cases, like the Deep Viewpoint system [
Cleaning. In computational data science, cleaning refers to filtering out unwanted data points. This notion relates well with the case of citizen contributed content who can be considered harmful, for example, for including aggressive or hateful speech, or for disclosing personal information. Citizen contributed content is monitored, moderated, and possibly removed.
Analysis. Analyses can be quantitative, such as statistical methods to visitors’ responses. More complex analytics involve automatic or semi-automatic content analysis algorithms (e.g. to identify content that requires monitoring, e.g. detecting hate speech), or automated reasoning (e.g. for emotion classification [23]). All these methods somehow augment the data. Content analyses in citizen curation applications span different methods but the most striking difference with data science is the role of human intervention. This is not surprising but opens the opportunity of expanding data journeys to cases where the workflow is more complex than a input/output data flow, and multiple user interventions occur during the process. Crucially, new input in the form of annotations can be produced, in search for similarities or differences between citizen responses (compare) and are typically mediated by user interfaces. However, qualitative analyses may be also performed offline with the aid of questionnaires, surveys, or in focus groups.
Movement. Citizen curation applications transfer data across different systems. In SPICE, these include one (or more) engagement interfaces (e.g. via tablets or mobile phones) often sustained by Web APIs of intermediate, “headless” applications which, in turn, access catalogue metadata collected and augmented on a common data hub. Such data hub acts as mediator between application and cultural heritage archival systems and takes care of publishing the collections’ digital items (and their metadata) with the applications and, in turn, support the sharing of such metadata across applications, and the sharing of collected responses with the museums archive [24].
Reuse. In citizen curation applications, the most widely form or reuse relates to cultural heritage digital artefacts being picked by museum professionals at the beginning of the process (select). Digitization processes can be also considered forms of reuse. Visitor responses can be analysed by dashboards and other tools that are separated from the systems where they originate from.
Visualisation. This activity is the one that differs significantly from stand-alone data science pipelines to more interactive, user-intensive applications in this domain. Users of all types (curators, citizens, researchers) interact with the data in many ways, engaging with the digital objects via viewing, listening, and sometimes touching, or immersing themselves in a virtual or augmented reality. Citizens engage with the artifacts via mediated experiences that involve in-presence storytelling as well as hybrid modalities, for example, combining physical reconstructions of objects (e.g. a wooden or 3D-printed replica) with digital representations in cyber-physical spaces.
Another set of classes in DJO refer to roles that data can have: input, output, temporary, etc… (see Fig. 2) These are certainly valid in any software engineering setting, but they don’t tell much about the types of resources (assets) involved. Citizen curation data journeys require a tailored characterisation of the types of (digital) objects involved, that can be in-turn mapped to several different nodes in the data flow. The concept of citizen curation object is used here to refer to any digital resource used or created through the citizen curation process. It includes (i) digital representations of artworks and their metadata and museum labels, (ii) resources that guide the citizen curation activity (e.g. quizzes, interpretation exercises) and (iii) the results of the activity (e.g. citizen answers, stories, interpretations). The data and metadata associated with the results of citizen curation activities (e.g. citizen answers to questions plus metadata associated with the author (e.g. their identity and community membership) and content of the activity (e.g. the text and extracted features such as its sentiment and values) is what is essential to making sense of these journeys. Table 2 shows a list of resource types derived from analysing the five SPICE case studies (but many more could emerge in a broader survey). Dilemma (Prompt), Stories (Autoethnographies), Photos Artwork metadata, Artwork images, Interview, Survey, Question, Prompt, Stories (Autoethnographies), Scripts, Response, Text, Choice
Images, Essay, Game, Puzzle
We have seen how the backbone components of computational data journeys are compatible with citizen curation applications. We also observed how the ontology would require additional components to express the richness of these systems, both in terms of activities and types of resources involved. However, to reflect further on the opportunities derived from such a perspective it is worth considering the key issue of capturing data journeys and reflecting on the additional representational layers needed.
In the case of the SPICE Linked Data Hub, the project developed an activity monitoring layer that has the purpose of recording events from connected citizen curation applications, linking catalogued artifacts with citizen responses, and make them reusable for analysis. The backbone representational layer is the established W3C Prov-O data model [25]. The model can be further extended covering the specificity of citizen curation artifacts and activities. It is worth noting how such representational layer is agnostic with respect to the underlying technology. Events described as such could be stored in a traditional relational database, in a graph database, or a blockchain [26].
Artworks, metadata, and responses travel across various systems whose competences vary from hosting data, monitoring activities, and providing extended user interfaces. Citizen responses may be generated by users interacting with a mobile application, when the underlying system (the app itself) generates a new event referencing the artifact, the activity performed, and the response. End-user systems can be seen as operating in a diversified social media network. Another citizen may receive a notification, via another citizen engagement system in SPICE, asking to react to a newly generated response. The new user comments with an emoticon, and the underlying infrastructure record the new event.
An extended set of activities and object types as well as tracing user operations in event graphs can help in answering the issues mentioned before, for the benefit of both citizens and museum practitioners.
Sense-making. By representing citizen curation applications as data journeys, we can potentially support analytics covering multiple dimensions of interest to museum professionals and researchers, including emotions, preferences, and choices, and finding similarities and differences across communities of users. Curators can use the data journeys to explore the responses to a given artifact in specific interaction contexts, responses of a given community, or how different activities relate. Data journeys can be leveraged by an analytics dashboard able to support curators in exploring the contributions from a multiplicity of perspectives, independently from the diversity of user-facing systems. Event graphs allow for analysing users’ behaviours and contextualising the content within a rich interaction context. Data journeys can be leveraged for observing reception by difference community groups, sensitivities, and cultural backgrounds, and explore ways of characterising it in the archive without imposing one view over others.
Monitoring sensitive content. New content produced by end-user systems can be monitored thanks to the events graph before it reaches the archive. Data managers (curators, developers) can review the collected information, as well as establish moderation processes or make use of intelligent systems for content analysis, to automatically flag content that can be inappropriate or potentially violate the privacy of users. Some content can be flagged as sensitive but still considered of value to be included in the archive to preserve the authenticity of the original response.
Ownership. Requesting a specific citizen response, the citizen curation application will be able to also know who produced that response, whether it was authored and by whom, accessing the full lineage of the resource. Similarly, when a citizen curation script uses an artifact image, the data journey will describe how that metadata record was created, the link to the original source, being it the museums’ Website or a collection management system. Citizens can claim ownership of contributed content and ask for specific attribution statements to be included.
Terms of use. Data journeys provide a high-level representation of how a certain asset (artifact, image, etc...) is being used, by whom, and for what purpose, in end-user applications. Such representation could be fine-grained and leverage existing standards for digital rights management such as W3C ODRL [27]. Information can include the usage policies applicable to that context so that applications can adapt and mediate intelligently with their users, relying on a standard protocol for adapting user interfaces to content, such as IIIF [28].
We can summarise what has been discussed so far about citizen curation data journeys in five layers:
Resources: resources used in the data journey such as artwork images, metadata records, data sources, licencing information, and terms of use, each one identified by a Linked Data entity URI, and leveraging a rich set of types (extending Table 2).
Event graphs: events occurring in different systems, such as a citizen curation activity that generates a user response or a curator who selects a response to be included in the archive. Data nodes graph: a graph of data-to-data relationships, such as reused tools and resources manipulated by the process, abstracting from the event graphs, that characterise the data flow of a given application, focusing on its design (abstracting on actual events).
Activity graph: a graph of high-level activities. In the context of citizen curation, these can be specialisations of the general scenario introduced before (but expanding on the activities listed in Table 1).
Policies graph: a graph of metadata about ownership, licences, and terms of use, to support the computational analysis of terms of use.
Before concluding, it can be useful to reflect on the implications of having these different layers altogether and on possible, concrete use cases that can benefit from such a holistic view.
Tracing sensitivity. Heritage institutions are typically strongly characterised both geographically and culturally. Some institutions may find sensitive certain content because of how it triggers local, unsolved issues. Other content can be considered sensitive by museum curators at a certain point in time but this may change in the future. The heritage archive is the perfect place where this knowledge can be preserved. Crucially, data journeys may support the study of how reception changes with changing cultural norms.
Granularity of policies. Licences (terms of use) may be associated to a whole collection of entities (such as catalogue metadata). However, catalogue-level terms of use may not be applicable to all items in the same way, as specific photograph of an artwork, for example, may have special ownership and terms of use. Linked data and related graph technologies can help in specifying terms with a high degree of granularity, pointing to collections, items, or their parts.
Composite objects. Citizen curation applications may generate composite objects, including images of artefacts, curators’ notes (e.g. questions of a slow looking activity), and citizen contributed content. Data journeys allow to capture information about rights and terms of use and reason upon the compatibility of rights when joining content in composite objects.
Compatibility of terms of use. Applications should make users aware of the difference in terms of use associated with each one of them. Potentially, an intelligent system could raise issues in relation to intended use (using a simulated workflow to verify agreement with current policies).
Visibility and access control. When terms of use affect access control, relevant users should be notified and instructed on what type of actions are needed to ensure a continued availability of resources.
Time-dependent information. It is not uncommon for museums to have limited control on the terms and conditions associated with the artworks, and often negotiate with artists (or rights management agencies) terms bound to specific contexts (e.g. a festival). Data journeys allows to capture and preserve the usage policies, ownership and terms of use associated with assets involved in a specific citizen curation scenario at the time of the events occurred.
Changes of terms of use. When an owner changes the terms of use of an image, they should be notified that there are applications having rights to access that image for a purpose that should not be allowed anymore. In this case, the owner may decide to either revoke the permission or restore the original policies.
Revoke consent. Similarly, curators shall know if a citizen does not want their content to be used anymore, and such changes should be propagated to citizen curation applications.
In this article, we argued that a knowledge organisation system for “data journeys”, such as the DJO, can help in disentangling problems that include issues of sense-making, ownership, sensitivity, privacy, and rights management. Data journeys can help govern the complexity of citizen curation applications. Here, we observed citizen curation by abstracting it as a data-intensive system. However, to realise this vision, many problems need to be solved concerning infrastructure, technologies, and methods that could support the implementation of data journeys. To that end, stakeholders should privilege open standards and distribution rather than offering end-to-end systems in isolated silos.
The research leading to this publication has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements "SPICE - Social Cohesion, Participation, and Inclusion through Cultural Engagement" (Grant Agreement N. 870811), https://spiceh2020.eu, and "Polifonia: a digital harmoniser of musical cultural heritage" (Grant Agreement N. 101004746), https://polifonia-project.eu. The publication reflects the author’s views. The Research Executive Agency (REA) is not liable for any use that may be made of the information contained therein. [15] E. Daga, L. Asprino, R. Damiano, M. Daquino, B. Diaz Agudo, A. Gangemi, T. Kuflik et al, Integrating citizen experiences in cultural heritage archives: requirements, state of the art, and challenges, ACM Journal on Computing and Cultural Heritage (JOCCH) 15, no. 1 (2022) 1-35. [16] G. Chowdhury, From digital libraries to digital preservation research: the importance of users and context, Journal of documentation 66 (2010). doi:10.1108/00220411011023625. [17] L. Moreau, The foundations for provenance on the web, Now Publishers Inc, 2010. [18] A. Harth, A. Polleres, S. Decker. Towards a social provenance model for the web, Workshop on
Principles of Provenance (PrOPr), 2007. [19] S. Leonelli, N. Tempini, Data journeys in the sciences, Springer Nature, 2020. [20] E. Daga, P. Groth, Data journeys: explaining AI workflows through abstraction, Semantic Web (2023) 1-27. [21] E. Daga, M. d'Aquin, A. Gangemi, E. Motta, Propagation of policies in rich data flows, in:
Proceedings of the 8th International Conference on Knowledge Capture, New York, 2015, pp. 1-8. [22] D. Garijo, P. Alper, K. Belhajjame, O. Corcho, W. Gil, C. Goble, Common motifs in scientific workflows: An empirical analysis, Future Generation Computers Systems 36, (2014) 338-341. [23] A. Lieto, M. Striani, C. Gena, E. Dolza, A.M. Marras, G. Luca Pozzato, R. Damiano, A sensemaking system for grouping and suggesting stories from multiple affective viewpoints in museums, arXiv preprint arXiv:2304.14117, 2023. [24] J. Carvalho, E. Daga, L. Asprino, A. Gangemi, M. Maguire, P. Mulholland, A. Stoneman, C.
Uwasomba, SPICE Linked Data Hub-Data infrastructure and tools to support Citizen Curation, Curation, in: ExICE - Extended Intelligence for Cultural Engagement, Bologna, 2023. [25] WC3, PROV-O: The PROV Ontology, 2013. URL: https://www.w3.org/TR/prov-o/ [26] G. Anagnostakis, Towards a blockchain architecture for cultural heritage tokens, in: Transdisciplinary Multispectral Modeling and Cooperation for the Preservation of Cultural Heritage, First International Conference, TMM_CH 2018, Athens, Greece, October 10–13, 2018, pp. 541-551. [27] I. Renato, Open digital rights language (ODRL), in: Open Content Licensing: Cultivating the
Creative Commons, Sydney University Press, 2007. [28] S. Snydman, R. Sanderson, T. Cramer, The International Image Interoperability Framework (IIIF): A community & technology approach for web-based images, in: Archiving conference, vol. 2015, Society for Imaging Science and Technology, pp. 16-21.