D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy

Towards a sustainable Open Data ECOsystem D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 955569. The opinions expressed in this document reflect only the author’s view and in no way reflect the European Commission’s opinions. The European Commission is not responsible for any use that may be made of the information it contains. An approach to steer the behaviour of non-government data holders towards open data through a technical strategy Project Acronym ODECO Project Title Towards a sustainable Open Data ECOsystem Grant Agreement No. 955569 Start date of Project 01-10-2021 Duration of the Project 48 months Deliverable Number D4.2 Deliverable Title An approach to steer the behaviour of nongovernment data holders towards open data through a technical strategy Dissemination Level Public Deliverable Leader University of the Aegean (UAEGEAN) Submission Date 22-10-2024 Author (s) Charalampos Alexopoulos - UAEGEAN, Maria Ioanna Maratsi - UAEGEAN, Mohsan Ali – UAEGEAN, Georgios Papageorgiou - Farosnet S.A., Abdul Aziz – UNIZAR, Dagoberto Herrera Murillo - UNIZAR Document history Version # Date Description (Section, page number) Author & Organisation V0.0 06-06-2024 Kick-off Meeting Charalampos Alexopoulos - UAEGEAN V0.1 05-06-2024 First draft table of contents Charalampos Alexopoulos - UAEGEAN, Maria Ioanna Maratsi - UAEGEAN V0.1a 06-06-2024 Review of table of content Andrea Polini - UNICAM, Charalampos Alexopoulos - UAEGEAN, Maria Ioanna Maratsi - UAEGEAN, Mohsan Ali – UAEGEAN, Georgios Papageorgiou - Farosnet S.A., Abdul Aziz – UNIZAR, Dagoberto Herrera Murillo - UNIZAR V0.2 20-08-2024 First draft Charalampos Alexopoulos - UAEGEAN, Maria Ioanna Maratsi - UAEGEAN, Mohsan Ali – UAEGEAN, Georgios Papageorgiou - Farosnet S.A., Abdul Aziz – UNIZAR, Dagoberto Herrera Murillo - UNIZAR V0.2a 13-09-2024 Second draft. Restructuration after the feedback from the first peer-review. Charalampos Alexopoulos - UAEGEAN, Maria Ioanna Maratsi - UAEGEAN, Mohsan Ali – UAEGEAN, Georgios Papageorgiou - Farosnet S.A., Abdul Aziz – UNIZAR, Dagoberto Herrera Murillo – UNIZAR D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy Version # Date Description (Section, page number) Author & Organisation Reviewers: Javier Nogueras - UNIZAR, Francisco Javier López Pellicer - UNIZAR V0.3 18-10-2024 Final report. Editorial work after the feedback from the second peer-review Reviewers: Bastiaan van Loenen - TUDELFT, Javier Nogueras, UNIZAR, Francisco Javier López Pellicer - UNIZAR, Birger Larsen - AAU V1.0 22-10-2024 Final editing. Danitsja van Heusden-van Winden – TUDELFT D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy Table of Contents Abbreviations ..................................................................................................................................................................... 7 1. Introduction ............................................................................................................................................................... 9 1.1 Problem Definition .................................................................................................................................... 9 1.2 Role of this deliverable in the ODECO project .............................................................................10 1.3 Structure ......................................................................................................................................................10 2. Methodology ..........................................................................................................................................................11 3. Barriers and Motivations for Data Sharing for Non-Governmental Data Holder Groups ........13 3.1 Non-Specialist Citizens ..........................................................................................................................13 3.2 Data Journalists .........................................................................................................................................13 3.3 Elementary School Students ................................................................................................................15 3.4 Non-Governmental Organisations ....................................................................................................16 3.5 Commercial Organisations ...................................................................................................................17 3.6 Open Data Intermediaries ....................................................................................................................18 4. Technical Requirements for Non-Government Data Holder Groups ...............................................20 4.1 Open Data Action Principles ................................................................................................................20 4.2 Technical Requirements for OD Holders and Users ...................................................................21 4.3 Synthesis ......................................................................................................................................................24 5. Challenges for OD Creation and Sharing ....................................................................................................28 5.1 Enhanced data creation process through better quality, integration and validation tools 28 5.2 Enhanced data sharing through better licensing and anonymisation ................................29 5.3 Towards inclusive, user-driven and circular open data platforms through better interfaces and feedback mechanisms ................................................................................................................29 6. Technical Steering Mechanisms per Challenge ........................................................................................30 6.1 Existing relevant solutions ....................................................................................................................30 6.1.1 General toolkits for open data sharing ......................................................................................30 6.1.2 Integration of non-government data on open data repositories ...................................33 6.1.3 Open data generated by volunteers through micro tasking: the case of the HOT Tasking Manager ...................................................................................................................................................35 6.2 Proposed technical solutions within the ODECO Project .........................................................38 6.2.1 Proposed solutions for Challenge 1 - Enhanced data creation process through better quality, integration and validation tools ......................................................................................................38 6.2.2 Proposed solutions for Challenge 2 - Enhanced data sharing through better licensing and anonymisation ............................................................................................................................47 6.2.3 Proposed solutions for Challenge 3 - Towards inclusive, user-driven and circular open data platforms through better interfaces and feedback mechanisms .................................50 6.3 Technology Readinessof the Proposed Solutions .......................................................................58 7 Discussion and Conclusion ................................................................................................................................59 7.1 Shared perspectives ................................................................................................................................59 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 7.2 Limitations ...................................................................................................................................................60 7.3 Future research directions and actions ...........................................................................................60 Acknowledgements .......................................................................................................................................................61 Appendix A – The ODECO TW5 Ideathon Structure ........................................................................................63 Appendix B – User-centred Design .........................................................................................................................65 References .........................................................................................................................................................................68 List of figures Figure 1: The methodological process of Task 4.2 ............................................................................................11 Figure 2: Main components of Epicollect5 ...........................................................................................................31 Figure 3: Key attributes of QField. ...........................................................................................................................32 Figure 4: Screenshot of the process of ingesting individual datasets into the French open data portal ...................................................................................................................................................................................34 Figure 5: Screenshot of the dashboard of the datasets ingested by a user on the French open data portal ...................................................................................................................................................................................34 Figure 6: The project tasks screen in HOT-TM displays key information in two sections. On the left sidebar, users can find project details, task statuses in list form, step-by-step task instructions, and contributor statistics......................................................................................................................................................35 Figure 7: Screenshot of the mapping phase in HOT-TM. Source: https://tasks.hotosm.org/projects/15476/tasks .................................................................................................36 Figure 8: HOT-TM API that allows access to data on campaigns, projects, organizations, tasks, comments, among others. Source: https://tasks.hotosm.org/api-docs ...................................................37 Figure 9: A simple example of different data schema use in a Cultural Heritage case ......................39 Figure 10: The proposed implementation process ...........................................................................................40 Figure 11: Indicative screen prototype of the mapping recommendation upon dataset upload..41 Figure 12: Proposed methodology for the automated thematic annotation of open datasets .....42 Figure 13: BRYT Keyword Extraction Architecture .............................................................................................43 Figure 14: Example Scenario for the Application ...............................................................................................44 Figure 15: Minimum viable product for Application functionalities and working ................................44 Figure 16: The technology stack required to develop the Application .....................................................45 Figure 17: Building blocks of the Open Data Driven Citizen Science Applications generation model ...............................................................................................................................................................................................46 Figure 18: The development process of the proposed platform ................................................................48 Figure 19: Key components of the system ...........................................................................................................49 Figure 20: Three-part frame of the proposed user interface ........................................................................49 Figure 21: The development process of the proposed platform ................................................................50 Figure 22: An example of a Chatbot assistant on the Greek open data portal .....................................51 Figure 23: Channel 1 for a multi-channel inclusive feedback mechanism for diverse communities ...............................................................................................................................................................................................55 Figure 24: Channel 2 for a multi-channel inclusive feedback mechanism for diverse communities ...............................................................................................................................................................................................55 Figure 25: Example case of interaction with a dataset ....................................................................................56 Figure 26: An example of a user interaction with a dataset on data.gouv.fr ..........................................57 Figure 27: The user-centred design process based on ISO 9241-210:2010 ...........................................65 List of tables Table 1: Barriers to non-expert citizens’ contribution to OD ecosystems ...............................................13 Table 2: Motivations for non-expert citizens to contribute to OD ecosystems.....................................13 Table 3: Identified barriers ..........................................................................................................................................17 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy Table 4: Motivations for NPOs to contribute open data ................................................................................17 Table 5: Barriers of commercial organisations to contribute open data ..................................................18 Table 6: Motivations for commercial organisations to contribute open data .......................................18 Table 7: Barriers for open data intermediaries to contribute open data .................................................19 Table 8: Motivations for open data intermediaries to contribute open data .........................................19 Table 9 - Technical Requirements for NGO Open Data users (taken from ODECO deliverable D2.2) ...............................................................................................................................................................................................22 Table 10: Technical requirements, action principles, barriers, and motivations per Open Data Holder Group ...................................................................................................................................................................25 Table 11: Evaluation of Open Data Tools for the Key Challenges...............................................................32 Table 12: Technology Readiness of the Proposed Solution ..........................................................................58 Table 13 - GenAI tools for user interface design ...............................................................................................67 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 7 Abbreviations AI Artificial Intelligence API Application Programming Interface AR Augmented Reality D Deliverable ESR Early-Stage Researcher GDPR General Data Protection Regulation GenAI Generative Artificial Intelligence HOT Humanitarian OpenStreetMap Team HOT-TM Humanitarian OpenStreetMap Team Tasking Manager IDIP Intelligent Data Interaction Platform IOT Internet of Things LLM Large Language Model LODC Linked Open Data Cloud MS Milestone MVP Minimum Viable Product NGD Non-Government Data NGO Non-Governmental Organisation NLP Natural Language Processing NPO Non-Profit Organisation OD Open Data ODP Open Data Portal ODK Open Data Kit ODECO Open Data ECOsystem OSM OpenStreetMap PII Personally Identifiable Information RAG Retrieval-augmented generation TAM Technology Acceptance Model TRL Technology Readiness Levels TSLO Technical, Semantic, Legal, and Organizational VR Virtual Reality TW Training Week UCD User-centred design VGI Voluntary Geographic Information WAI Web Accessibility Initiative WP Work Package Nr Partner Partner short name Country Beneficiary 1 Technische Universiteit Delft TU Delft Netherlands 2 Katholieke Universiteit Leuven KUL Belgium 3 Centre National de la Recherche Scientifique CNRS France 4 Universidad de Zaragoza UNIZAR Spain 5 Panepistimio Aigaiou UAEGEAN Greece 6 Aalborg Universitet AAU Denmark 7 Università degli Studi di Camerino UNICAM Italy 8 Farosnet S.A. FAROSNET S.A. Greece D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 8 Nr Partner Partner short name Country Partner organisations 1 7eData 7EDATA Spain 2 Digitaal Vlaanderen DV Belgium 3 City of Copenhagen COP Denmark 4 City of Rotterdam RDAM Netherlands 5 CoC Playful Minds CoC Denmark 6 Derilinx DERI Ireland 7 ESRI ESRI Netherlands 8 Maggioli S.p.A MAG Italy 9 National Centre of Geographic Information CNIG Spain 10 Open Knowledge Belgium OKB Belgium 11 SWECO SWECO Netherlands 12 The government lab GLAB United States of America 13 Agency for Data Supply and Infrastructure ADSI Denmark 14 GFOSS Open Technologies Alliance GFOSS Greece 15 Inno3 Consulting IC France 16 Regione Marche RM Italy 17 Open Data Institute OCI United Kingdom 18 Swedish National Archives (Riksarkivet) SwNA Sweden D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 9 1. Introduction 1.1 Problem Definition Open Data initiatives often focus on a narrow selection of providers and user groups while neglecting other stakeholders in the ecosystem. This exclusion is evident in the dominance of governmental actors in providing open data. Sustainable open data ecosystems advocate for inclusivity, suggesting that non-government actors should also share their data in open formats (Charalabidis et al., 2018, Runeson et al., 2021). However, encouraging non-government actors to share their data as open data remains a largely unexplored challenge. The ODECO project aims to address this issue through a multidisciplinary approach that includes: (1) understanding the barriers and motivations for non-government actors to contribute open data to the ecosystem, (2) developing the technical mechanisms to facilitate this process, and (3) establishing governance structures to support and sustain it. Specifically, Task 4.2 examines the technical mechanisms that can be used to encourage nongovernment data holders to share open data. Task 4.2 considers the barriers and motivations of task 4.1, and technical requirements for non-government actors to become active contributors to the open data ecosystem (by releasing their data into the open data ecosystem) and proposes the corresponding technical steering mechanisms to suitably prepare the ground towards this direction. The technical steering mechanisms presented in this Deliverable are of varying Technology Readiness Levels (TRL), those developed within the ODECO project being of higher TRL. Governmental actors have historically been the dominant providers of open data. In many cases, such responsibility is engraved in policy documents, from non-binding strategic plans to binding laws and regulations such as the EU Open Data Directive. On the other hand, while there have been some initiatives by non-governmental actors to publish their data as open data, it is still a limited practice and remains largely desired (van Loenen et al., 2018). Non-governmental actors are actors that are outside the public/governmental sector, such as companies, civil society organisations, and the media. Exploring the motivations and challenges for non-governmental actors to contribute open data is thus essential to develop sustainable open data ecosystems that incorporate both government and non-government open data. Aligned with the ODECO’s Description of Action, this deliverable aims to answer the following main objective: • Identify and further develop technical mechanisms (solutions) to facilitate OD sharing by non-government data holder groups. These solutions will entail quality assurance methods, the development (or embedment) of data integrator services, as well as new user (data holder) interfaces from at least the working prototypes of the engaged ESRs. These solutions will help the non-governmental data holders to more easily create and share their data as open data using these novel and automated solutions. First and foremost, the developed solutions could be embedded to the already existing open data spaces (e.g. Zenodo, OpenAIRE, wikidata etc.) that enable the upload of datasets from any non-governmental stakeholder. Secondly, they could be used towards the (further) development of their own open data portals (or websites sharing open data) within their organisations. In order to address the main objective, three sub-objectives are necessary: • Sub-objective 1: Analyse the identified motivations and technical requirements for nongovernmental data holders to contribute open data. The identified motivations are briefly presented in section 3 of this deliverable and in more detail in D4.1. • Sub-objective 2: Analyse the technical requirements and barriers which hinder nongovernmental data holders to contribute open data. The developed and proposed technical mechanisms should not disregard the requirements of the data users to build “user-driven” D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 16 Motivations • Being active citizens. Celis Vargas et al. (2023) have identified that specially OD learning activities seeking the development of competencies for active citizenship address activities for collection of own data. In those cases, students have been involved in both creating simple spreadsheets and collecting more complex data by using tools such as sensors, games, or mobile applications (Badioze Zaman et al., 2021; Chicaiza et al., 2017; Saddiqa, Larsen, et al., 2019; Vallejo-Figueroa et al., 2018). The motivation is actively participating as citizens to create a better world. For example, students during the focus interviews wondered about their school projects "How is this going to create a better world?". • Raising awareness of local issues around students' context and daily life. Considering students as experts of their own local experience, they can potentially create and share local datasets addressing aspects of their environment and daily life experiences. From their perspective, they want to raise awareness and provide contextual understanding of local issues. Their motivation behind is raising their voice and being heard. Making "Children's voice as important as others". • Helping the community around the school, considering students as important actors in local ecosystems. Students are motivated by helping the community around them, implicitly for a feeling of belonging, building their identity and their place. Creating and sharing data has been identified as an opportunity to also building networks in their local communities by addressing problems from other actors and contributing to solve them with data. • Seeing what students learn in schools as useful in the real world. For young pupils in elementary school, it was relevant to see what they do in school being used in the real world. It increases the authenticity of their learning experience. • Making school activities more relevant, interesting, and fun. To sum up, creating and sharing their data might increases their motivation for learning by fulfilling their intrinsic motivation for making something relevant, being heard and connecting to their communities. Overall, students are motivated by active learning experiences, were they have freedom to experiment and learn by themselves. 3.4 Non-Governmental Organisations Non-Governmental Organisations (NGOs), also interchangeably called Non-Profit Organisations (NPOs) in this section, take up an intermediary role in the open data ecosystem, where they bridge the gap between open data providers and users (González-Zapata and Heeks, 2015). NPOs are unique as intermediaries because there are specific user communities they are focusing on to address a social issue (Enaholo, 2017) while also not seeking to gain any profits from it (Salamon and Anheier, 1992). As such, NGOs can be seen as groups of individual users that are with one more level of explainability and technical assistance in the use of available tools. The difference with non-specialist citizens is that the services and guidelines could be drafted in a more technical format. If the aim of NPO is to improve overall openness and transparency, it can push them to aim for a variety of projects and have open data and open source on that principle (Baack, 2015). However, NGOs may face barriers that prevent them from contributing (Chattapadhyay, 2014). The identified barriers and motivations are presented in Table 3 and Table 4 respectively. The barriers found are summarised in Table 3 ; there are two barriers to the motivations that stop NGOs from sharing the data as open: (1) the lack of financial resources, limits NGOs ability to pursue open data projects, (2) the lack of knowledge and technical skills of NGOs’ employees, affecting the NGOs capacity to correctly license and publish dataset, and (3) the lack of an existing common portal for NGOs to share their data. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 17 Table 3: Identified barriers Barriers Description The lack of financial resources The nature of non-profit organisational model means that NGOs need to secure funding in a form of grants or individual donations. Some NGOs may not be able to obtain enough funding for their open data projects. The lack of knowledge and technical skills of NGOs’ employees The lack of knowledge and technical skills of NGOs’ employees prevents them from publishing the data they have as open, or they do not know how to provide it properly licensed as open. NGOs that are short on their financial resources are unable to hire additional employees with the required skills. The lack of an existing common portal for NGOs to share their data NGOs cannot add to the open government data portals and there is no existing common portal for NGOs and civil societies. However, creating such a platform and/or maintaining it would require a lot of financial and human resources that the average NGO cannot afford. Table 4: Motivations for NPOs to contribute open data Motivation Description Show the power of open data Creating a project with enhanced or reused open data can engage other stakeholders or make them more aware of the open data available. Help with the local and global societal issues There is open government data that might be available on societal issues, but not utilised. By finding ways to enhance the data and publish it as open, the NPO can highlight it and help those affected by the issue. Receive positive feedback from the community The community can give feedback on the issues with the data and give suggestions when they are interested and involved which motivates NGOs to continue the project. Follow organisational goals and the personal beliefs of the employees in openness and transparency Some NPOs have transparency and openness of data and knowledge as their main goals and employees join NPOs because they share the vision. Create opportunities for other stakeholders NGOs want their data to be reused widely by and benefit a variety of communities, NGOs, private and governmental organisations. 3.5 Commercial Organisations Commercial organisations are defined as those whose goal is to make an economic profit. This is illustrated in contrast to users who intend non-commercial use, which Creative Commons (Creative Commons, 2023) defines as “means not primarily intended for or directed towards commercial advantage or monetary compensation.” Commercial organisations may have different motivations to contribute to OD ecosystems, depending on their field and own necessities. The identified barriers and motivations for commercial organisations to share data are presented in Table 5 and Table 6 respectively. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 18 Table 5: Barriers of commercial organisations to contribute open data Barrier Description Technical/ Tools Lack of data integration tools, and tools for importing large datasets. License compatibility Datasets generated by commercial organisations are sometimes a combined product of different sources, for which the license may not be compatible. Insufficient resources [SMEs] Limited money, labour and time, compared to large organisations. Resistance by other community members [Big corporations] Resistance by community members as they fear big corporations may dominate the project landscape. Table 6: Motivations for commercial organisations to contribute open data Motivation Description Improving the data quality By improving the data quality organisations can deliver better services to their customers. Community building Some organisations are interested in each topic or domain. They kickstart communities in that topic or domain to maintain data continuity and freshness. Standardization and alignment with an open project schema For some organisations, this is a motivation to contribute data to open projects , to lower the subjectivity of the data. Being grateful and wanting to give back Some organisations feel that open projects have provided value to them and their products, and feel it is only fair to give back and provide value to the project, to keep it going. 3.6 Open Data Intermediaries Open data intermediaries are defined as “third-party actors who provide specialised resources and capabilities to (i) enhance the supply, flow, and/or use of open data and/or (ii) strengthen the relationships among various open data stakeholders” (Shaharudin et al., 2023). Examples are developers who process and include open data in apps/software, crowdsourcing platforms that gather and publish data as open data, and organisations that transform open data into easily digestible information such as visual forms. Open Data Intermediaries carry out a wide range of tasks depending on their specialised resources and capabilities such as compiling data, validating data, and improving technical openness of data (Shaharudin et al., 2023). Thus, through some of their tasks, some open data intermediaries (re-)produce data that could be contributed back to the open data ecosystem. As all actors in the open data ecosystem, including open data intermediaries, have their own agency, and thus self-interested (Davies, 2011; Poikola et al., 2010), the motivations for them to contribute open data naturally have to be aligned with their interests. In other words, open data intermediaries either must be convinced that they can directly or indirectly capture value for themselves by publishing open data (intrinsic motivation) or they must be forced by external conditions such as through law and regulations to do so (extrinsic motivation). The identified barriers and motivations for open data intermediaries are presented in Table 7 and Table 8 respectively. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 19 Table 7: Barriers for open data intermediaries to contribute open data Barriers Description Protecting business interests Some open data intermediaries are hesitant to release (some of) their data as open data because they want to protect their business interests. No compelling legislation There are limited or no legal requirements that compel open data intermediaries to provide open data. Additional costs to develop and maintain open data platforms Developing and maintaining data platforms to release open data would incur additional costs to open data intermediaries. Table 8: Motivations for open data intermediaries to contribute open data Motivation Description Support the visibility of their organizations By providing open data, open data intermediaries can increase the visibility of their organization; hence, it is a form of marketing for their products and services. Support other partners within their networks Some open data intermediaries provide open data to support the business or operation of their partners, which they would also get the benefit from. This may further strengthen their position within their network. The desire to contribute to society Some open data intermediaries are driven by philanthropic or altruistic motivations to contribute open data that could benefit society. The availability of open data platforms to share open data Open data intermediaries could be more motivated to release open data if there are open data platforms that could facilitate them to do so. This is because not all open data intermediaries have the capability and resources to develop and maintain their own data platforms. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 20 4. Technical Requirements for Non-Government Data Holder Groups For many non-government actors studied, motivations related to own benefit were identified in D4.1. Such motivations include supporting other partners within an organization network for their own benefit, the private value of contributing, the feeling of belonging, and enjoyment. Aligning these own benefit motivations of the different stakeholders is key to achieving the goal of open data sharing. These own benefits match with the intrinsic motivations from the Self Determination Theory defined in Task 3.3 (D3.3). Own benefit motivations are supported by the desire to create an impact, mentioned in the clusters of supporting other partners, supporting the community, and creating social impact. These can be classified as extrinsic motivations in the Self Determination Theory. Finally, the availability of and desire to improve the technical skills and solutions are mentioned as motivations, and potential motivations. This means that creating the correct technical environment can be an enabler for open data sharing, as well as including the potential stakeholders in the process, with the desire to improve the technical environment. 4.1 Open Data Action Principles Based on the findings of ODECO Task 3.3, where the commons-based governance model for the open data ecosystem has been refined, six action principles have been unveiled related to boundary-making, community support, participation and decision-making, legal mechanisms, interoperability and interoperable projects, and sustainability. In our analysis, we have focused on the following principles which are the most relevant to the non-governmental data holders: 1. Encouraging participation and shared decision-making, 2. Considering appropriate legal mechanisms, 3. Designing an ecology of interoperable projects, and 4. Ensuring sustainability of open data ecosystems. The aim is to extract the main key points and propose the development and develop technical solutions that can motivate each group to be an active participant in the open data ecosystem. Encouraging participation and shared decision-making The key point of this action principle is the opportunity for less powerful users of the open data ecosystem (particularly the ones that in a hierarchical system have no control over the open data) to raise concerns, and needs, and provide feedback. It is important to mention that for public and non-profit organizations, technical solutions may motivate them to give voice to citizens and underrepresented groups, but for commercial users (intermediaries, companies, etc.), they must have profit as an incentive or other extrinsic motivations (laws, regulations, etc.). Journalists, NGOs, local governments, and open data intermediaries can all benefit from collaborative platforms that enable public contribution and feedback on open data. For journalists, such platforms offer valuable crowdsourced insights, though robust moderation is crucial to maintain data integrity. NGOs require simplified, low-maintenance feedback mechanisms to efficiently gather input without straining resources. Local governments can leverage citizen feedback to improve services and decision-making. Open data intermediaries play a key role by providing technical solutions for these feedback systems, bridging the gap between data providers and users. While students may not directly implement technical solutions, they can draw inspiration from these key points to understand the importance of public engagement in D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 21 open data ecosystems. Ultimately, a well-designed feedback mechanism can enhance data quality, increase transparency, and foster greater civic participation across various sectors. Considering appropriate legal mechanisms Two integral issues were uncovered in this legal mechanism action principle. The predominant one revolves around the lack of data literacy and the impact that legal mechanisms can have, especially in government structures. The second impediment is the discovery and adoption of sustainable business models for intermediaries and commercial users. This obstacle involves finding economically viable ways for these entities to operate while adhering to open data principles and legal requirements. The primary challenge faced by all users is the requirement for compatible licenses. A potential solution to this issue could be the development of an ontological model for open data licensing. This model could be applied to open data sets, thereby automating the process of checking compatibility between different licenses. Designing an ecology of interoperable projects The integral focus of this action principle is the adoption of interoperable standards between datasets so they can be exchanged, combined, and analysed even when they are sourced by different organizations. This initiative can boost the use of open data across various user types as it will significantly reduce the technical skills and expertise currently required for users to effectively utilize the open data they can acquire and use. The most crucial problems mentioned in this action point in D3.3 are, as already noted, the lack of personnel with the appropriate skill set and knowledge to utilize open data. Secondly, there are concerns among users regarding potential security and privacy issues that may arise from combining different datasets, which could indirectly expose data that was obscured through anonymization techniques in the original datasets. Ensuring sustainability of open data ecosystems The primary emphasis of this action point is to ensure the sustainability of the open data ecosystem and the continued participation of users within it. This mainly involves the struggle for funds and resources, particularly for non-profit organizations. The focus is on advocacy for public funds, while for commercial users, the emphasis is on the viability of business models that centre on active participation in the open data ecosystem. 4.2 Technical Requirements for OD Holders and Users In this Section, the technical requirements for Open Data holders were mainly derived from ODECO Task 2.2 (D.2.2) and were mapped afterwards with the data holder group they concern. Analysing the requirements regarding their importance (number of NGD user group reporting this requirement) and connection to the NGD holders (authors’ interpretation), the most relevant requirements derived from D2.2 are those related to the Reusability and Interoperability aspects. They are mapped regarding their respective stakeholder group shown in Table 9. Since, NGD holders cannot share data through the open government data portals, we specifically mention the already existing open data spaces (e.g. Zenodo, OpenAIRE, wikidata etc.) that enable the upload of datasets from any non-governmental stakeholder. In this case, the available tools and registries for NGD holders should comply to the FAIR principles. As already mentioned, the developed and proposed technical mechanisms should not disregard the requirements of the data users in order to build “user-driven” solutions. Furthermore, to achieve circularity in the open data ecosystem, we need to consider the dual role of the nongovernmental stakeholders of the ecosystem serving both as users and providers of open data. This means that specific open data user requirements are also important for the development of technical mechanism to stimulate NGD Holders to provide their data as open data. For instance, one requirement from the user perspective is something that needs to be resolved by the provider D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 22 or data holder. In this case, we use these requirements to identify the challenges that occurred for the NGD data holders. Table 9 - Technical Requirements for NGO Open Data users (taken from ODECO deliverable D2.2) Requirements Nonspeciali st actors Data journalis ts Element ary school student s Nongovern mental organisa tions Comme rcial organis ations Open data interm ediarie s R1 - Facilities for data processing and for checking the quality of data X X X X X R2 - Datasets should have schema.org tags to provide a semantic meaning X X X R3 - Visualize a dataset on the data portal X X X X X R4 - Data provenance X X X R5 - Technical standards for data formats, metadata, APIs X X X X R6 - Easily publish data in the required formats X X X X X X R7 - Engage with data holders to understand their technical needs and challenges X X X X X X R8 - Feedback mechanisms for data holders to suggest improvements X X X X X X R9 - Reducing the gap between NGO user groups and their technological needs can help identify the common requirements of all stakeholders X X X X X X R10 - High quality data provision X X X X X X R11 - The development of data integrators to enable the technical interoperability of open data stemming from different domains X X X X X R12 - Widely adoptable data formats X X X X X X R13 - Metadata suitability checker for data publishers X X X X X X R14 - Quality enhancement tools (data augmentation tools) should be available X R15 - Users should be allowed to upload updated versions of existing datasets X X X X X X R16 - Consistent and reliable delivery of datasets should be provided to the artificial users X D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 23 These requirements related mainly to reusability and interoperability are described in more detail as follows: User needs for Reusability for an Inclusive Open Data Ecosystem from a technical perspective • R1: Facilities for data processing and for checking the quality of data: Currently, portal administrators and data users must have the basic skills for data processing and for checking the quality of data. For open data reusability, data quality, validity and completeness of the data may be evaluated with the use of technologies available via the data portal. This requirement addresses the services that need to be developed to reduce the data processing capabilities of the portal administrators. • R2: Datasets should have schema.org tags to provide a semantic meaning: All the datasets available on the open data portal, even third-party datasets, should have schema.org tags to provide a semantic meaning because with this requirement we can prioritize in searches the comments that can be more relevant to their profile. It can also help in reusing the dataset. • R3: Visualize a dataset on the data portal: If a user wants to visualize a dataset on the data portal, then the data portal must have visualisation tools and resources. For better insights of the dataset, it will be better to visualize it in the portal before downloading and processing it for further tasks. • R4: Data provenance: Data provenance with respect to the reusability of open data within open data portals ensures data authenticity and quality. This involves clearly documenting the dataset’s origin, ownership, and updates. Metadata and reuse standards like open data licenses promote data sharing and accessibility. • R5: Technical standards for data formats, metadata, APIs: Implementing CSV, JSON, and XML data formats provides platform portability. Searching for data and context comprehension improve significantly by using Dublin Core or schema.org metadata standards. Strong APIs like REST or GraphQL provide automated data access, integration, and scalability while retaining security and control. Data is readily accessible, interpreted, and used by varied people and systems because of these standards. • R6: Easily publish data in the required formats: An Open Data Ecosystem requires data to be publishable in the right formats. Data vendors may alter and disseminate their data in CSV, JSON, and XML using straightforward tools and services that simplify data transformation and publication. This streamlines the process, lowers technical barriers, and ensures data format compliance, improving end-user accessibility and usability. • R7: Engage with data holders to understand their technical needs and challenges: Engaging data holders to understand their technological requirements and difficulties is essential for an inclusive Open Data Ecosystem. Conducting surveys, seminars, and feedback sessions helps data providers determine their needs and challenges. This knowledge enables the creation of customized assistance, training, and technological solutions to help data holders publish and distribute high-quality data. • R8: Feedback mechanisms for data holders to suggest improvements: Continuously improving the Open Data Ecosystem requires data holders to provide input. Discussion forums, feedback forms, and writing to data publishers enable data suppliers to share their knowledge. Feedback is crucial for improving data standards and tools, ensuring the open data ecosystem meets users demands, and identifying areas for development. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 24 User needs for Technical Interoperability for an Inclusive Open Data Ecosystem from a Technical Perspective Technical requirements outlined in Table 10 from ODECO Deliverable 2.2, relevant to the open data users from a technical standpoint, are presented. Some parts of these requirements are listed below with their explanation, related to the non-government stakeholders: • R9: Reducing the gap between NGO user groups and their technological needs can help identify the common requirements of all stakeholders. A one-size-fits-all technological approach would be more beneficial for stakeholders from government or non-government sectors. The solutions should be ready to enhance both types of platforms (national or governmental data portals but also other open data portals like Zenodo and OpenAire). • R10: High quality data provision – From the government to the non-government stakeholders, and vice versa. The creation of open data has three different perspectives: (a) the actual creation of data (e.g. from IoT sensors), (b) the creation of DCAT-complied datasets (flat and contextual metadata levels) and (c) the linked open datasets (detailed metadata level) creation. Solutions should target all of the above. • R11: The development of data integrators to enable the technical interoperability of open data stemming from different domains. In this stream, non-government stakeholders can integrate their data in a better way. The proposed solutions should target easier (automated if possible) creation of data and metadata from their systems. • R12: Widely adoptable data formats – deciding upon uniform adoptable data formats can resolve several problems in the current open data ecosystems and can make open data interoperability much easier. • R13: Metadata suitability checker for data publishers – metadata selection is a very important phase of the open data publication. For non-governmental organizations, there is also a need for a metadata suitability checker before adding the data to main streams of open data portals (National or governmental open data portals). • R14: Quality enhancement tools (data augmentation tools) should be available – With the rise of AI and LLMs, data augmentation tools can enhance the quality of open data. Leveraging AI to achieve semantic and syntactic interoperability is highly sought after in the market. An example of this scenario is AI for data and data for AI. • R15: Users should be allowed to upload updated versions of existing datasets. It has been noted that users currently cannot upload their improved data back to the portal, whether the improvements pertain to data quality, metadata, or applications. • R16: Consistent and reliable delivery of datasets should be provided to the artificial users – Is there a way to bridge the gap between LLMs and open data portals? By establishing a technical interoperability channel, LLMs can effectively utilize data from open data portals. 4.3 Synthesis This subsection includes the summary of the technical requirements (D2.2), action principles (D3.3), barriers, and motivations (D4.1) for non-governmental data holder groups to share their data as open data, grouping the various elements according to an experts’ grouping as shown in Table 10. As explained in the methodology Section, the synthesis of the Challenges was conducted by the core authors team of this Deliverable. The Challenges and their detailed description are presented in the next Chapter (Chapter 5). D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 25 Table 10: Technical requirements, action principles, barriers, and motivations per Open Data Holder Group Requirements + Action Principles /NonGovernmental Data Holder Group Challenge Group NonSpecialist Citizens Data Journalists NonGovernme ntal Organisa - tions Schools Commerci al Organisa - tions Open Data Intermedi aries Importanc e count Easily publish data in the required formats 1 X X X X X X 6 The interoperability needs of a government body are different from the interoperability needs of a company (fragmented open data ecosystems) 1 X X X X X X 6 Useful data discovery tools (e.g., Contextualized discovery) 1 X X X X X X 6 High quality data provision 1 X X X X X X 6 The ODP should provide technical assistance for the contributor to prepare the data resource to be added to the portal (data preparation, cleaning, labelling) 1 X X X X X X 6 Widely adoptable data formats 1 X X X X X X 6 The ODP should help the data contributor discover useful connections from the data resources it contributes to other resources already present in the portal 1 X X X X X X 6 Basic skills for data processing and for checking the quality of data. 1 X X X X X 5 The development of data integrators to enable the technical interoperability of open data stemming from different domains 1 X X X X X 5 Technical standards for data formats, metadata, APIs. 1 X X X X 4 Datasets should have schema.org tags to provide a semantic meaning. 1 X X X 3 Addressing technical data challenges 1 X X 2 Face challenges with data quality, timeliness, and resources 1 X 1 Benefit from seamless data exchange 1 X 1 Face challenges adopting new interoperable formats 1 X 1 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 32 • QField QField is a mobile application that is meant to integrate perfectly with QGIS, which is a popular desktop GIS software program. Its purpose is to collect geographical data. In addition to providing a variety of geographic data gathering and editing operations directly on mobile devices, it enables users to take their QGIS projects out into the field that they have created. The following Figure 3 highlights the key attributes of QField: Figure 3: Key attributes of QField. Beside the key attributes here are a few use cases: • Urban planning and infrastructure mapping. • Agricultural data collection and farm management. Below, Table 11 compares the four data collecting and publishing tools (Datasette, Epicollect5, ODK, and QField) to the specific user requirements mentioned earlier. Table 11 also presents a technical assessment of each tool in correspondence to the Challenge it addresses: Table 11: Evaluation of Open Data Tools for the Key Challenges Challenge Datasette Epicollect5 ODK QField Challenge 1: Enhanced data creation process through better quality, integration and validation tools Supports basic SQL queries for data validation and analysis Basic data validation through form design and constraints Supports data validation rules and constraints in forms. Supports exports to CSV, JSON, XML Relies on QGIS for data validation, some in-app validation tools Challenge 2: Enhanced data sharing through better licensing and anonymisation Metadata fields for licensing, requires custom anonymization No built-in support, export for external anonymization Metadata support, potential for anonymization through form design Managed through QGIS metadata and external anonymization Challenge 3: Towards inclusive, user-driven and circular open data platforms through better interfaces and feedback mechanisms GitHub issues, community forums, Built-in visualization tools Community feedback through app, Basic visualization for data collection Feedback via community channels, surveys, Limited in-app visualization, better with external tools GitHub, QGIS forums for feedback, Limited in-app, relies on QGIS for advanced visualization D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 33 6.1.2 Integration of non-government data on open data repositories Open Data portals are online platforms that enable the discovery, access, and management of metadata records for datasets (Kubler et al., 2018). These datasets may be available for download in various formats. These portals play a crucial role in making data accessible to the public. In the European Union, data.europa.eu is the leading open data initiative that brings together datasets from community providers at all levels of government (Carrara et al., 2015). In turn, all member states have a national open data portal (Herrera-Murillo et al., 2022). Current user interfaces in open data ecosystems do not facilitate the addition of data by external actors outside government data providers (van Loenen et al., 2021). The latest edition of the Open Data Maturity Report (Publications Office of the European Union, 2023) offers several recommendations for leveraging collective intelligence by engaging the broader open data community in national open data programs. Key suggestions include allowing re-users to upload their own data and showcase their ideas and creations on national portals. Additionally, the report advocates for enabling users to comment on and rate datasets, with these inputs being integrated into search algorithms. We conducted a review of the European Open Data Portal and 27 national portals, as compiled by Herrera-Murillo et al. (Herrera-Murillo et al., 2022), totalling 28 portals. During this inspection, we specifically assessed whether these portals explicitly and visibly offer users the following features: a) the ability to request the addition of specific datasets, b) the option to submit usergenerated use cases or applications, and c) the opportunity for non-governmental users to contribute their data to the portal. Out of the 28 portals reviewed, 17 (61%) provide users with the option to request the addition of datasets. However, the implementation of this feature varies significantly across portals. In some instances, a generic help request form is used, where users select the category of their request, with dataset incorporation requests being one of the options. Other portals offer more sophisticated mechanisms; for example, the Estonian portal displays all received dataset proposals and allows users to vote on their favourites, while the Lithuanian portal presents a request history that includes the outcomes—whether favourable or unfavourable—along with explanations for each decision. Eighteen of the 28 portals (64%) allow users to submit applications or use cases to be showcased on the site. However, the methods for submission vary widely. Some portals simply invite users to share their use cases via email, offering little guidance on the required structure. In contrast, other portals provide a dedicated form for submissions. A notable example is the Austrian portal, which features a comprehensive form with multiple fields, including title, description, nature of the application, open datasets used, links to the application, links to the source code, and administrator information. Only 6 of the 28 portals (21%) allow providers outside the public administration to ingest their own data into the portal. These countries are France, Spain, Austria, Poland, Luxembourg and Portugal. The portals of France, Austria, Luxembourg, and Portugal permit uploads from all types of users. The Polish portal extends this capability to companies. In contrast, Spain restricts the uploading of non-governmental data to select private institutions, primarily foundations, associations, and professional bodies that generate data from publicly funded projects. To illustrate how a non-governmental provider can add its data to an open data portal, we will use the French portal as a reference. The portal offers a range of resources detailing best practices for open data publishing, as well as specific guidelines on how the platform operates. Data providers begin by registering an account, which allows them to add datasets. The process starts D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 34 with including the metadata for the resource, including the title, description, license, update frequency, keywords, time coverage, and spatial granularity. After this, the file itself is uploaded (See Figure 4). Finally, the publisher is encouraged to share the newly created resource. Once the dataset is published and accessible to the public, the data provider gains access to a dashboard where they can edit the entry and monitor various metrics related to the dataset's usage and interaction. These metrics include views, subscribers, use cases, comments, community contributions, quality indicators and recommendations to improve data and metadata quality (See Figure 5). Non-government providers can also use the portal API to automate data ingestion into the portal. Figure 4: Screenshot of the process of ingesting individual datasets into the French open data portal Figure 5: Screenshot of the dashboard of the datasets ingested by a user on the French open data portal The addition of third-party datasets is not limited to government open data portals; there are also numerous open-access repositories that enable a wide range of users, particularly researchers, to share datasets generated from their work and research processes. A notable example is the Zenodo repository, managed by CERN, which currently allows the creation of entries containing D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 35 up to 100 files and 50 gigabytes of data. These entries can include both final and intermediate research outputs, such as datasets. Other general repositories include Figshare, Dryad, and the Open Science Framework. In addition, there are domain-specific repositories, such as ICPSR for social sciences, as well as institutional repositories hosted by various academic and research institutions. 6.1.3 Open data generated by volunteers through micro tasking: the case of the HOT Tasking Manager This section reviews the main lessons learned from the ODECO project research on user interfaces for open data generation, specifically in the context of humanitarian mapping Herrera-Murillo et al. (2024). Voluntary Geographic Information (VGI) is a philosophy that harnesses the collaborative contributions of volunteers to collect, analyse and distribute geographic data. This is one of the most successful examples of open data contribution by non-governmental actors. VGI plays with the idea of “citizens as sensor” which appeals to the ability of human beings to synthesize and interpret local information in an intelligent way (Goodchild, 2007). In turn, the Humanitarian OpenStreetMap Team (HOT) is one of the most recognized VGI initiatives. It leverages the OpenStreetMap (OSM) platform to enable the development of freely accessible global maps, with a special emphasis on regions in critical need of humanitarian assistance. A relevant tool in this process is the HOT Tasking Manager (HOT-TM), which coordinates a worldwide network of volunteers to support focused and efficient mapping efforts. Figure 6 and Figure 7 shows the project tasks screen where volunteers select the mapping task to be performed and the screen corresponding to an individual task. Figure 6: The project tasks screen in HOT-TM displays key information in two sections. On the left sidebar, users can find project details, task statuses in list form, step-by-step task instructions, and contributor statistics D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 36 Figure 7: Screenshot of the mapping phase in HOT-TM. Source: https://tasks.hotosm.org/projects/15476/tasks As part of the ODECO project, Herrera-Murillo et al. (2024) conducted a systematic analysis of the humanitarian mapping process within HOT-TM, aiming to understand the interactions between volunteers and the user interface of the platform. Based on this research, the authors developed a set of practical recommendations for designing platforms and user interfaces that facilitate data contributions from non-governmental users, especially those that seek to involve the public, regardless of whether geographic information is involved: Technology is a powerful enabler for building a community of data contributors: HOT-TM serves as a model that can be replicated by other non-governmental open data initiatives. Its technological platform acts as a bridge between a robust volunteer community, like OSM, and humanitarian organizations responding to global disasters. The platform equips both groups with the necessary user interface, tools and information to create and validate open data. Volunteer experience is a key asset: when volunteers sign up for HOT-TM, their previous mapping experience within OpenStreetMap is automatically shared through indicators. This enables project managers to set experience requirements for participation in projects based on their complexity. The task validation phase is reserved for more experienced users, who possess the necessary judgment to evaluate and correct mapping products effectively. Research shows that this distinction is well-founded, as advanced users consistently demonstrate higher productivity, better task success rates, and sustained participation over time. Open data initiatives by non-governmental actors can greatly benefit from identifying key variables to assess data contributor experience. Doing so allows for more efficient task assignment, ensuring that projects are matched with the right level of expertise. Managing the expertise of data contributors also involves addressing the challenge of retaining and motivating novice users, many of whom stop contributing after their first project. This suggests that initiatives should focus on flattening the learning curve for newcomers, providing them with constructive feedback and friendly user interfaces, and sharing indicators of the real-world impact of their contributions. Micro tasking can simplify participation for data contributors: HOT-TM employs a micro tasking approach that, unlike regular OSM dynamics, directs users toward specific, simple mapping objectives. Most tasks can be completed in just a few minutes, allowing even inexperienced users to contribute meaningfully to the project goals. This micro tasking model can be applied to other non-governmental open data aggregation initiatives, where participants are asked to make small and manageable contributions. When combined, these individual efforts can have a significant impact, making large-scale projects more accessible and efficient through the power of crowdsourcing. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 37 Data quality matters: a common concern with open data generated by non-governmental stakeholders is whether it meets the necessary quality standards for future use. While many of these stakeholders possess valuable data, they often lack the skills and tools for effective data and metadata quality management. Overburdening them with additional quality control tasks could discourage their participation or divert their focus. To address this, HOT-TM has developed a twostage quality control mechanism. In the first stage, during the mapping phase, contributors focus on achieving the mapping objectives as quickly as possible, ensuring that geographic information for emergency response is available immediately. In the second stage, during the validation phase, project organizers and experienced contributors validate and correct the mapping tasks, ensuring that the final product meets quality standards. This approach demonstrates how any open data initiative involving non-governmental actors can benefit from similar quality control mechanisms, whether through automated rule-based checks or reviews by human experts. Process data is a useful tool for continuous improvement: HOT-TM collects detailed data on contributors, the work process, and the resulting outputs. This information is shared with relevant communities via a well-documented API (see Figure 8). This approach enables data-driven management of non-governmental open data initiatives, allowing external contributors to engage in process improvement within an open innovation framework. Key improvement decisions, such as those related to user interface design, can be effectively informed by analysing interaction data collected from users. Figure 8: HOT-TM API that allows access to data on campaigns, projects, organizations, tasks, comments, among others. Source: https://tasks.hotosm.org/api-docs Harnessing the collective intelligence of non-governmental open data contributors: Research indicates that HOT-TM is highly efficient in achieving broad mapping objectives within a short timeframe. However, its success relies heavily on experienced mappers, while novice mappers—who make up most contributors—tend to have a lesser impact, partly due to suboptimal interaction with the platform. As platforms like HOT-TM develop more sophisticated mechanisms for aggregating user input, they can tap into the "wisdom of the crowds," where the collective contributions of a diverse group result in an outcome of higher quality than the sum of individual inputs. After identifying the tools for data sharing described above, it is noticeable that these tools allow for the data creation and sharing types of platforms, however, the need to make the created datasets more accessible in general open data portals remains. Many of the technical steering D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 38 mechanisms (or solutions) proposed in the following Section, are not limited to (even though they could improve those as well) government data sources (e.g., government open data portals), but they are applicable for use on other non-governmental data sources and providers, which is mainly the focus of this Deliverable. 6.2 Proposed technical solutions within the ODECO Project In this Section, proposed technical solutions corresponding to each of the Challenges described in Chapter 5 are presented. The proposed solutions are either a result of ongoing research within the ODECO project (thus not finalised but under progress), or a result of a proposed solution from the ODECO TW5 Ideathon. An important aspect to consider is that there is a distinction between the technical solutions within the ODECO project and the ones proposed during the OD Ideathon in Samos. The former ones are ongoing research of the ODECO team, and they will continue to progress while the project is still ongoing, but the latter are solution ideas to be considered and might be (usually the case) at a much earlier stage of conception, whose progress into deeper levels of progress is out of the scope of this deliverable. Another important note for the following Section of this document is the distinction between the term “user” in its conventional use and meaning (so anyone, including citizens, non-specialist user groups etc.) and the use for this deliverable. In this Section, whenever the term “user” is encountered, it is not limited to the user as simply a data user, but rather, it refers to a system user, so the target system user who will make use of the proposed technical solutions presented. In our case, and as the main purpose of this deliverable, this of course refers to the nongovernment data holder groups which we aim to motivate to share their data, using the proposed and envisioned technical solutions. 6.2.1 Proposed solutions for Challenge 1 - Enhanced data creation process through better quality, integration and validation tools 6.2.1.1 Augmenting semantic interoperability for datasets This proposed semantic interoperability solution aims to target one of the common problems that arises when non-government data must be combined and integrated with other the data sources: the semantic heterogeneity between the schemas (conceptual models) of two data sources that must be combined. The main challenge related to semantic interoperability that organisations and data portals face is related to the inconsistent use of data vocabularies and metadata schemas (Maratsi et al., 2024 a); Organizations use different (custom) data and metadata schemas that do not match existing standardised vocabularies and ontologies to express domain-specific knowledge. For instance, taking an example from the Cultural Heritage domain, and assuming we have two organisations (e.g., art gallery, museum, etc.), the artifacts exhibited in Art Gallery 1 and Art Gallery 2 of Figure 9, are not automatically interoperable when exchanging data from these two institutions, due to the lack of linking the custom schemas to known, standardised ones of this specific domain (e.g., CIDOC-CRM and more, in this case). The quest for standardisation and interlinking existing ontologies and schemas mainly in the cultural heritage domain, but also in other domains (e.g., legal domain, psychology), has been actively researched within the ODECO project, the results of this study to be found in (Maratsi et al., 2024 b, Loutsaris et al., 2023, Maratsi et al., 2024 c, Maratsi et al., 2023). D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 39 Figure 9: A simple example of different data schema use in a Cultural Heritage case If the custom schema fields of the two different institutions could be mapped in an automated or semi-automated way to existing, standard schemas upon dataset upload for instance, then the new data would instantly become semantically interoperable with more data sources, and more easily discoverable by semantic search engines. The proposed solution aims to use an LLM (e.g., LLAMA) to align and map these various schemas to standardized, vocabularies. This model would be trained using expert-curated data to ensure accuracy and would follow the process of Figure 10. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 40 Figure 10: The proposed implementation process Initially, the LLM model will be trained on curated expert-validated data (examples of experts’ mappings), and then will go through a preliminary system validation process which will be supervised (approved or rejected) by a human. This is going to be an iterative process until consensus is reached regarding the produced mappings to standardised vocabularies and schemas. On a similar note, the use of Linked Open Vocabularies, for the facilitation of this process of ontology and standard domain-specific vocabulary re-use for improved semantic interoperability is described by (Maratsi et al., 2024 d). Some technologies to be used for the implementation of this solution are the following: • Data Ingestion & Storage: AWS S3 • Model Training and Deployment: o Model Training: PyTorch, TensorFlow o Training Environment: GPUs and TPUs on AWS SageMaker • Model Deployment: Docker • Data Integration: Apache Kafka • Monitoring & Logging: Prometheus The outcome of this process is a coherent data integration which ensures all data is aligned with the proposed system's standard vocabularies. The form of outcome is a produced recommendation of mappings to standardised representation schemas of a specific domain (or more generic if this is not necessary). The results of this process will then be contributed to the Linked Open Data Cloud (LODC). In practice, when an organisation aims to share their data (e.g., upload to an OD portal) or elsewhere, the produced system solution will provide them with a set of recommendations on how their custom schema can be improved in terms of semantic interoperability. Figure 11 shows an indicative screen prototype of the mapping recommendation process to the end user (e.g., organisation or data holder group which wants to share their data). D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 41 Figure 11: Indicative screen prototype of the mapping recommendation upon dataset upload This proposed technological solution can be utilised, for instance, as a SaaS data standardisation platform, opening various opportunities to be integrated in existing open data portals, or as a separate service. This solution aims to improve the status on semantic interoperability of datasets and representation schemas, but also consequently improves dataset findability, while at the same time makes active contributions to the Linked Open Data paradigm. 6.2.1.2 Automatic Thematic Annotation of Open Data Whenever a non-government data asset is integrated in an open data portal, it is essential to classify and annotate it appropriately to facilitate its findability and access by potential users. This problem is similar to the problem faced by Open Data portals such as the European Data portal (data.europa.eu) that must aggregate and classify the contents harvested from heterogeneous national and local portals to predefined list of themes. Although ideally the original datasets should be already annotated with a theme vocabulary and later matched with the European data themes, based on our desktop review of the datasets, it is evident that assigned themes are often either missing or incorrect. Incorrect theme assignment can hinder the findability of the datasets by other stakeholders. To address these issues, we developed a solution that assigns proper themes to datasets, even when a dataset belongs to more than one category. This solution is a thematic classification system designed to resolve the problem of incorrect thematic assignment. We created a multiclass classification model using a training dataset of 29,793 records in RDF format. The system employs both classification techniques and cosine similarity, analysing the metadata (title and description) of datasets against the themes defined by the European Open Data Portal. The developed system can assign themes to datasets based on their title and description with 92% accuracy. Additionally, the model can suggest a second theme that could be assigned to a dataset, improving its findability on the portal. All the code and data are available on the github repository. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 48 Figure 18 shows the proposed solution development process to create a first Minimum Viable Product (MVP) of the provided service: Figure 18: The development process of the proposed platform This proposed solution reduces the risk of data exposure during development and testing phases, protecting sensitive information while it also improves data quality by ensuring high data quality and integrity, leading to better decision-making and operational efficiency. Moreover, it simplifies compliance with data-related regulations by establishing clear policies and procedures for data management, integrating data licensing sand compliance under the same platform. 6.2.2.2. An Open License Decision Support Software One of the problems that system users (in this case, including interested non-government data holder groups to share data) are encountering is the complexity of licensing currently used in open data sets. Now, there are more than 300 open data licenses in use. Choosing the right license requires specialized knowledge to understand the various licenses accompanying open data sets and to properly reuse them. It becomes even more complicated in cases where users must combine data sets, especially from different providers and from different countries with different legal frameworks. This can be a severe impediment and discourages open data users not only from using open data but, more importantly, from becoming active contributors to the open data ecosystem. Therefore, the aim of this proposed solution is the creation of an open license decision support system that can help open data creators and users to handle the difficult decision of picking a data license. Although there are some tools that provide this functionality, some drawbacks have been detected. The design of these tools is static – they either provide far too much information in a tabular form, or they exist as multi-step questionnaires where the respondent has no idea about their progress in the questionnaire or about how the questionnaire itself is structured. In this proposed solution, the goal is to develop a tool that can (Figure 19): • Provide a decision-support system to dataset creators to choose the right license according to their specific needs. • Build a system that will serve as a learning tool for dataset creators on how to navigate the different considerations when they must select a data license. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 49 Figure 19: Key components of the system First, the team aims to design a simple questionnaire to be answered by dataset creators or reusers. This questionnaire will prompt respondents to reflect on the source and provenance of each individual data element in their dataset, their objectives for creating and sharing the dataset, and their proprietary or other interests in the dataset. Each response will be mapped to nodes within the backend decision tree, generating a list of licenses that meet the respondent's parameters. Second, the team aims for respondents to be able to "see" the internal workings of the decisionsupport software. This feature enables the software to serve as a learning tool for respondents as well. It also allows for more transparency of the software and moves away from a technosolutionist "black box" design. The goal is to create a user interface where the respondent can always see each section of the questionnaire as well as their progress. The team also wants to enable respondents to toggle between questions and alter answers. Finally, they intend for the license decision tree to be dynamic for the respondent. In other words, as the respondent progresses through the questionnaire, the license decision tree will be generated with each response and modified with each alteration of a response. Instead of being visible to the respondent only at the end of the questionnaire, this dynamic license tree will always be visible as they go through the questionnaire, allowing them to understand the underlying logic of the recommendation system. Figure 20 and Figure 21 explain the key components of the system and three frame of the proposed user interface for the system respectively. Figure 20: Three-part frame of the proposed user interface D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 50 In Figure 21, the development process of the system is depicted. The aim of the team was to adopt a process for the creation of the decision system that would "design solutions 'with' stakeholders, and not 'for' them." Therefore, a circular design for the development was adopted so that the system could be adjusted based on the requirements and needs of the users. Figure 21: The development process of the proposed platform The proposed open license decision support system addresses the complex challenge of navigating the multitude of open data licenses, offering a user-friendly solution for both creators and users of open data. By combining a simple questionnaire with a dynamic, transparent interface, the system not only recommends appropriate licenses but also serves as an educational tool, empowering users to understand the decision-making process. 6.2.3 Proposed solutions for Challenge 3 - Towards inclusive, user-driven and circular open data platforms through better interfaces and feedback mechanisms Imagine an open data portal in which a variety of user groups, such as government agencies, researchers, developers, and organizations that represent civil society, are actively engaging with datasets. The feedback requirements of these user groups are diverse; for instance, developers are looking for more technical data formats, while researchers may seek improvements to metadata to facilitate data interpretation. Despite the availability of feedback channels such as discussion forums, emails, and surveys, there is frequently a lack of transparency regarding the processing and implementation of this input. Consequently, users are left with a lack of clarity regarding the enhancements or modifications that are the outcome of their contributions. A possible feedback scenario might involve a user accessing a climate dataset through the national OGD portal. After working with the data, they notice inconsistencies in the date formats and provide detailed feedback via the discussion forum or fill in the feedback form. However, due to the fragmented nature of feedback processing across different channels (emails, feedback form, social media, discussion forums), their input is not addressed promptly, or the portal lacks transparency regarding what actions are taken based on such feedback. As a result, the feedback loop is incomplete, hindering data reusability and trust among users. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 51 6.2.3.1. A ChatBot Guide for Open Data Portals The motivation behind this proposed solution is the need of citizens without technical skills on open data to access information, tools and data available, knowing where and how to find them. Citizens without technical skills might need open data for understanding an issue, for example, climate change, compare Open Data and creating a solution to solve a problem. The idea pertains to a Chatbot assistant which profiles the user towards the resources they need on the open data portal and externally. Figure 22 shows the screenshot of chatbot. Figure 22: An example of a Chatbot assistant on the Greek open data portal For instance, as also shown in Figure 22 the Chatbot could ask questions to understand the user needs and help navigate them accordingly through the available data. Below are some indicative questions the Chatbot could ask: • What is your problem? • Do you want me to suggest topics of interest? • What data are you looking for? • Do you need me to direct you to the tools and guides to help you analyse the data? • There exists research on this topic using this data. Do you want me to direct you to it? There can also be suggestions of questions that might be of interest to the user. The technologies to be used for the implementation of this proposal are mainly Generative AI and Natural Language Processing (NLP). Creating a database of existing tools/visualisations and data research on the portal or from external stakeholders could enhance the trustworthiness of organisations, while the information for the Chatbot can be verified and/or provided by a list of organisations that the data portal owners deem as trustful. The value of the proposed solution for citizens (users) is that it helps them gain understanding of a topic, while having data and research in the same place. The process of accessing and using data is facilitated. From the open data portals’ side: new data users are being attracted. In addition, missing data or components can be discovered, as well as tools and other resources currently not available on the portal. The value of this solution might also be addressed towards private and third parties, such as through the creation of a database of stakeholders (NGOs, NPOs, volunteering organisations and more). D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 52 The development process of the Chatbot would include the affixing of an appropriate license, the preparation of the database, tools and research available to use, the verification of sources from external stakeholders, the creation of the algorithm and models, the design of the interface and guide character, the automation of solutions, conducting the appropriate usability tests, and forming a human support group of experts that would be available upon request. 6.2.3.2. Enhancing better interfaces and user interaction under feedback supported by AI This proposed solution addresses the challenge of better interfaces and user interaction with open data, solutions focused on creating more intuitive, engaging, and accessible ways for users to interact with data. The proposed interface will be hosted by a local data portal and will be addressed to all citizens (including vulnerable groups). The rationale behind this is finding the appropriate mechanisms to support and develop more intuitive, AI-powered interfaces that allow users to query and interact with data in natural language. These interfaces could provide personalized recommendations, visualizations, and insights based on user queries, also allowing for AR (Augmented Reality) integration to create immersive data interaction experiences, where users can explore complex datasets in a 3D environment, making it easier to understand patterns and relationships. Furthermore, voice-activated assistants that guide users through datasets could be introduced, offering explanations, answering questions, and helping with data analysis tasks. More specifically, the key features of the IDIP (Interactive Data Intelligence Platform) include: 1. Interactive Data Visualizations o Dynamic Dashboards: Create customizable, interactive dashboards that allow users to visualize data in real-time. Users can drag and drop data elements to create their own visualizations (charts, graphs, maps) and explore the data dynamically. 2. Augmented Reality (AR) and Virtual Reality (VR) Integration o Interactive Data Models: In AR, users could interact with holographic data models using gestures or voice commands. This is particularly useful for education, allowing students and professionals to engage with data in a hands-on, intuitive manner. The VR employs 3D near-eye displays and pose tracking to give the user an immersive feel of a virtual world and therefore the users can have a futuristic idea engaging also with data. 3. Voice-Activated Assistants providing feedback o Personalized Data Assistant: Introduce a voice-activated assistant that helps users navigate the open data platform, answers their queries, provides summaries, and even performs data analysis tasks. The assistant could also learn from user interactions to provide more personalized experiences over time. o Hands-Free Interaction: This feature is especially valuable for users who need to multitask or for those with accessibility needs. The voice assistant could execute commands, retrieve data, and explain complex datasets without the need for manual input. The technologies the features will be based on include: • An LLM Fine tuning - providing an LLM with additional and specialized data to refine the training. • Retrieval-augmented generation (RAG) a specific architecture in which queries are “augmented” with additional and domain specific/contextual information. Implementing RAG in an LLM-based question answering system has two main benefits: It ensures that the model has access to the most current, reliable facts, and that users have access to the model’s sources, ensuring that its claims can be checked for accuracy and ultimately trusted. • Technology Acceptance Model (TAM) (perceived usefulness, perceived ease of use) D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 53 o Impact Tracking: Track the impact of feedback over time, showing how user suggestions have led to tangible improvements in data quality, usability, or features. This transparency encourages continued user participation. Other technology characteristics which need to be taken under consideration are: • Accessibility: Easily obtainable by everyone. • Licensing: No restrictions on usage. • Interoperability: Usable across different platforms and systems. The perceived benefits of such an interface for the users include: • Increased User Engagement: By providing more intuitive and interactive ways to engage with data, users are more likely to explore and utilize open data resources, leading to more informed decision-making and innovation. • Broadened Accessibility: The inclusion of accessibility features and multi-language support ensures that a wider range of users, regardless of their technical expertise or physical abilities, can effectively use the platform. • Enhanced Data Utilization: With tools that simplify data discovery, analysis, and visualization, users can more easily derive insights and value from open data, leading to greater use and impact of these resources. • Facilitated Collaboration: The collaborative features foster a community around data exploration and analysis, encouraging knowledge sharing and joint problem-solving. By implementing an Intelligent Data Interaction Platform (IDIP), open data becomes more accessible, engaging, and useful for a broader audience, ultimately enhancing the impact and value of open data initiatives. Development process parameters Regarding the development process of this proposed solution, it will need to include the following: • User Training and Onboarding: Provide tutorials or onboarding sessions to help users understand how to effectively use the feedback system. This could include demonstrations of how feedback leads to tangible improvements. • Privacy and Anonymity: Ensure that users can provide feedback anonymously if they choose, especially when dealing with sensitive data or issues. Privacy considerations should be paramount, particularly in handling personally identifiable information (PII). • Open Source and Community Collaboration: Consider making the platform open source to encourage community-driven development and innovation. This also allows for broader collaboration and integration with other open data initiatives. For Natural Language Processing (NLP) Interfaces: • Conversational Querying: Develop an NLP-based interface that allows users to interact with data using natural language (i.e. using BERT). • Contextual Suggestions: The interface could offer suggestions as users type, predicting their needs based on the context of their queries. This feature helps guide users, especially those who may not be familiar with complex data structures or querying languages. For Augmented Reality (AR) and Virtual Reality (VR) Integration: • 3D Data Exploration: Develop AR/VR environments where users can explore complex datasets in 3D space. For example, in a VR environment, users could "walk through" a city map that displays population density, pollution levels, and transportation networks in a visually immersive way. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 54 Implementation and Deployment: • User-Centric Design Process: Develop the platform with a user-centric approach, involving potential users in the design and testing phases to ensure the interface meets their needs and expectations. • Modular Architecture: Build the platform using a modular architecture that allows for easy updates and the integration of new features. This ensures the platform can evolve with user needs and technological advancements: ( open data -> LLM model (BERT) –> RAG technique -> user interface ) • Cloud-Based Infrastructure: Deploy the platform on a cloud-based infrastructure to ensure scalability, accessibility, and efficient handling of large datasets. Sustainability: • Ensuring that the rights will belong to the local government. • A company will take care of the maintenance of the product, making necessary monitoring and adaptations using future science innovations and developments (after a public tender), ensuring the continuation of the funding). • User interface and technical specs for to be kept to a certified repository (i.e. ZENODO) – Open Science EU. 6.2.3.3. A multi-channel inclusive feedback mechanism for diverse communities The ability to provide feedback and user participation is critical for creating a fair digital future in a data-driven society. So, we propose the idea about a multi-channel feedback mechanism to foster inclusivity and engagement in open data sharing. It aims to empower both non-technical citizens and people with technical skills, to provide valuable feedback through open data portals. There are two main channels for this idea: • First, we want to empower those who already have data literacy skills to provide their feedback through open data portals. • Second, we want to find new ways to connect and include citizens with the open data portals through intermediation by public libraries. Public libraries offer a wealth of services to citizens and are, by design, inclusive places. This idea proposes a transformed feedback loop that not only addresses the technical needs of open data portals but also strengthens community bonds, reduces digital inequities, and empowers citizens across all levels to participate in shaping the data landscape. It also has the potential to enhance transparency, elevate public trust, and ensure that data-driven governance is truly reflective of all voices. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 55 Figure 23 and Figure 24 show the workflow of the described two channels. Figure 23: Channel 1 for a multi-channel inclusive feedback mechanism for diverse communities Figure 24: Channel 2 for a multi-channel inclusive feedback mechanism for diverse communities 6.2.3.4. A Plug-in for Feedback Mechanisms for Existing Open Data Portals The motivation behind this idea is trifold. First, from the data users’ side, the difficulty of providing feedback on the portal (lack of commenting features, lack of reporting tools), as well as the lack of awareness of how or why to interact with the portal, as well as what the effect of this interaction might be. In addition, users might include unofficial data sources, or wrong data. From the data holders’ side, there is also little to no interaction with the data, and the risk of unofficial data sources being used is also high. The data also runs the risk of becoming easily outdated. From the data portal owners’ side, the mentioned issues result in low traffic, less interaction with the website, low revenue from possible advertisement and cooperations, as well as problems with impact reporting (if relevant). Figure 25 shows an example of a dataset interaction case at the European Data Portal. The current available interaction with the data is: download, metadata, and validation. This figure shows an example of the limited options for interactive processes regarding the in-word dataset each time. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 56 There is no easy way to add feedback for this dataset, give recommendations on what is needed, or see other, existing commentaries from other users for this dataset. Figure 25: Example case of interaction with a dataset The proposed solution requires that issues are integrated in the data portals, that users can add issues in a user-friendly way, and they can propose changes (in the form of pull requests), and that users are generally aware of these possible interactions. For instance, the French data portal, already implements a commenting system into the datasets, as shown in Figure 26. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 57 Figure 26: An example of a user interaction with a dataset on data.gouv.fr Here we can see that under the Tab “Discussions” users can comment or start a feedback thread for this dataset. Technologies which can be used for this purpose include the W3C Web Accessibility Initiative (WAI), which provides tools for disabled people (e.g., people with hearing problems, impaired vision and other), Udata (pypi.org, 2024), a customisable social platform for open data, VCS software, for dataset version control, git, and more. The solution proposed is a plugin for open data portals, improving user experience in an easy and efficient way of communication and interaction with data and the portal owners, as well as data quality for specific use cases. In addition, data holders (owners) can provide higher quality data for improved efficiency and/or internal operations and gain new ideas for data provision. Data portals will increase their user-friendliness, and the higher quality of datasets will result in more impact of the platform, something vital for improving the open data ecosystem. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 64 Ideathon Programme Day 1 • Introduction and challenges presentation • Teams’ formation and facilitation • Teamwork Day 2 • Mentoring • Teamwork • Final presentation D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 65 Appendix B – User-centred Design User-centred design (UCD) is an iterative design process in which designers prioritize users and their needs at every stage. In UCD, design teams engage users throughout the design process using various research and design techniques to create highly usable and accessible products. According to ISO 9241-210:2010 (See Figure 27), once a plan for the design process is established, four interconnected human-centred design activities are undertaken: Figure 27: The user-centred design process based on ISO 9241-210:2010 Understand and specify the context of use: This activity involves identifying and describing the characteristics of the target users and other relevant stakeholders. It includes outlining the users' goals and tasks, as well as understanding the organizational, technical, and physical environments that define the system's context of use. Specify user requirements: In this phase, user needs are identified, and functional and other requirements for the product are specified. This process considers the context of use and aligns with the system's business objectives. Produce design solutions to meet user requirements: Design solutions are developed based on the context of use description. This begins with designing user tasks, user-system interaction, and the user interface to meet user requirements. Design solutions become more tangible through scenarios, simulations, prototypes, or mock-ups. Evaluate the designs against requirements: User-centred evaluation is essential throughout the project's various stages, starting from an early stage. While the participation of end-users in the evaluation is highly desirable, alternative methods that do not require direct user involvement can be used when practical constraints prevent their participation. These activities iterate until the designed solution meets user requirements. Hybrid intelligence Hybrid intelligence is the result of the integration between human and machine intelligence. This approach enhances human capabilities in the user interface design process. In particular, D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 66 generative artificial intelligence (GenAI) is gaining importance. This type of artificial intelligence can create products in forms such as text, images and videos by using generative models, often in response to humanly interpretable commands that describe the task to be performed or the desired output. In 2023, a Spanish developer named Javi López created "Angry Pumpkins" (Marr, 2024), a parody of the popular game Angry Birds, entirely using artificial intelligence tools like Midjourney, DALLE 3, and GPT-4. The development process involved creating graphical elements such as backgrounds, characters, and objects, as well as generating the source code of the game. López employed an iterative process, crafting a series of prompts to achieve the desired features without writing a single line of code. This example demonstrates the potential of generative AI to reduce time and effort in developing complex digital products. This is just one example of how GenAI is disrupting practitioners. Similarly, the academic community has responded with an emerging but expanding body of research on the implications of this technology for the design process. The available research has explored both the positive and negative implications of introducing this new technology (Li et al., 2024). On the positive side, GenAI offers the potential to enhance the capabilities of design teams by reducing solution development time and enabling humans to focus on activities that add greater value and usability. However, if automation is not implemented correctly, it may lead to the loss of the human touch in final design products. In response, proposals such as that of Weisz et al. (2024) have emerged, emphasizing the need to establish design principles for GenAI applications. These authors propose six principles: Responsible Design: Ensure that the AI system addresses real user problems and minimizes potential harm to users. Design for Mental Models: Communicate effectively with users on how to interact with the AI system, considering their background and goals. Design for Appropriate Confidence and Reliability: Help users determine when to trust or question the AI system’s results by encouraging scepticism regarding quality issues, inaccuracies, biases, underrepresentation, and other potential problems. Design for Generative Variability: Assist users in managing the capability of generative models to produce multiple distinct and varied results. Design for Co-Creation: Enable users to influence the generative process and collaborate effectively with the AI system. Design for Imperfection: Help users understand and work with results that may not always meet their expectations. Stige et al. (2023) have explored the impact of artificial intelligence on the human-centred design activity cycle. In the context of the use understanding phase, AI is primarily used for the automatic summarization of research findings into artifacts commonly used in subsequent phases, such as user personas and user scenarios. During the requirements specification phase, interacting with a GenAI can assist the development team in refining the final set of system requirements. However, the application of GenAI is most concentrated in the production of design solutions to meet user requirements. In this stage, GenAI can generate prototypes that are iteratively modified to develop detailed solutions. GenAI can assist designers in progressing from low-fidelity solutions (e.g., sketches) to high-fidelity solutions (e.g., GUI code) or optimize their designs. Finally, in the evaluation phase, where designs are assessed against requirements, GenAI can automate the detection of common pitfalls, allowing the test team to focus on identifying major flaws. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 67 Bleichner and Hermansson (2023) and Ho et al. (2024) have examined the application of GenAI in user interface design and identified the main commercial generative assistance tools: Galileo-AI, Uizard, and Midjourney. Table 13 highlights their main features. However, it is important to note that this list is not exhaustive, given the rapid emergence of new tools and the constant evolution of existing ones. Table 13 - GenAI tools for user interface design Tool Description Uizard (uizard.io) Uizard is an AI-powered service designed to assist in creating user interfaces and interactive prototypes. In addition to generating interfaces from text prompts, Uizard can scan screenshots and wireframes, converting them into editable mock-ups with customizable components. Galileo AI (usegalileo.ai) Galileo AI, launched in 2023, is a tool for designing both web and mobile user interfaces based on text or image prompts. It generates high-fidelity mock-ups that can be exported as HTML code and edited in other interface design tools. Midjourney (midjourney.com) Midjourney is a general-purpose text-toimage model accessible via Discord. While it is primarily used for generating images from text prompts, its output can also serve as high-quality prototypes for user interface design when provided with well-crafted prompts. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 68 References Ahmed, U., Alexopoulos, C., Piangerelli, M., & Polini, A. (2024). BRYT: Automated keyword extraction for open datasets. Intelligent Systems with Applications, 23, 200421. https://doi.org/10.1016/j.iswa.2024.200421 Ali, M., Papageorgiou, G., Aziz, A., Loukis, E., Charalabidis, Y., & Lopez Pellicer, F., J., (2024). Towards the Development of Interoperable Open Data Ecosystems: Harnessing the Technical, Semantic, Legal, and Organizational (TSLO) Interoperability Framework. In Proceedings of the 25th Annual International Conference on Digital Government Research (dg.o '24). Association for Computing Machinery, New York, NY, USA, 909–919. https://doi.org/10.1145/3657054.3657160 Anderson, J., Sarkar, D., & Palen, L. (2019). Corporate Editors in the Evolving Landscape of OpenStreetMap. ISPRS International Journal of Geo-Information , 8 (5), Article 5. https://doi.org/10.3390/ijgi8050232 Atenas, J., Havemann, L., & Priego, E. (2015). Open Data as Open Educational Resources: Towards Transversal Skills and Global Citizenship. Open Praxis , 7 (4). https://doi.org/10.5944/openpraxis.7.4.233 Baack, S. (2015). Datafication and empowerment: How the open data movement re-articulates notions of democracy, participation, and journalism. Big Data & Society , 2 (2), 2053951715594634. https://doi.org/10.1177/2053951715594634 Bleichner, A., & Hermansson, N. (2023). Investigating the usefulness of a generative AI when designing user interfaces (Dissertation). https://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva504251 Boyles, J. L. (2020). Laboratories for news? Experimenting with journalism hackathons. Journalism , 21 (9), 1338–1354. https://doi.org/10.1177/1464884917737213 BRYT Keyword Extraction GitHub Repository: GitHub - umairahmedq/BRYT-AKE Cambridge Dictionary. (2024). Motivation . https://dictionary.cambridge.org/dictionary/english/motivation Carrara, W., Chan, Wae-San., Fischer, Sander., & Van-Steenbergen, E. Creating value through open data: Study on the impact of re-use of public data resources. European Commission, 2015.https://data.europa.eu/sites/default/files/edp_creating_value_through_open_data_0.pdf Celis Vargas, A., Magnussen, R., Mulder, I., & Larsen, B. (2023). Towards a framework for Open Data literacy in education: A systematic mapping review of Open Data skills and learning approaches. Interaction Design and Architecture(s) , 57 , 133–151. https://doi.org/10.55612/s5002-057-008 Charalabidis, Yannis & Zuiderwijk, Anneke & Alexopoulos, Charalampos & Janssen, Marijn & Lampoltshammer, Thomas & Ferro, Enrico. (2018). The World of Open Data - Concepts, Methods, Tools and Experiences. Chattapadhyay, S. (2014). Access and use of government data by research and advocacy organisations in India: A survey of (potential) open data ecosystem. Proceedings of the 8th International Conference on Theory and Practice of Electronic Governance , 361–364. https://doi.org/10.1145/2691195.2691262 Chicaiza, J., Piedra, N., Lopez-Vargas, J., & Tovar-Caro, E. (2017). Recommendation of open educational resources. An approach based on linked open data. 2017 IEEE Global Engineering Education Conference (EDUCON) , 1316–1321. https://doi.org/10.1109/EDUCON.2017.7943018 Concilio, G., & Mulder, I. (2018). Open4Citizens Policy brief . OPEN4CITIZENS - Empowering citizens to make meaningful use of open data. http://open4citizens.eu/wpcontent/uploads/2016/01/O4C_D4.8_30.06.2018_Final.pdf Cook, K., Cakirlar, C., Goddard, T., Demuth, R. C., & Wells, J. (2018). Teaching Open Science: Published Data and Digital Literacy in Archaeology Classrooms. Advances in Archaeological Practice , 6 (2), 144–156. https://doi.org/10.1017/aap.2018.5 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 69 Corbett, J., Templier, M., & Takeda, H. (2018, January 3). The Making of a “Top” Open Data City: A Case Study of Edmonton’s Open Data Initiative. Proceedings of the 51st Hawaii International Conference on System Sciences . https://doi.org/10.24251/HICSS.2018.308 Corcho, O., Jiménez Blanco, J., Morote, C., and Simperl, E. (2022). Data.europa.eu and Citizengenerated DataOpportunities and challenges associated to the inclusion of citizen-generated data in data.europa.eu. Luxembourg: Publications Office of the European Union. https://data.europa.eu/sites/default/files/report/data.europa.eu_Report_Citizengenerateddataondata_europa_eu.pdf Coughlan, T. (2020). The use of open data as a material for learning. Educational Technology Research and Development , 68 (1), 383–411. https://doi.org/10.1007/s11423-019-09706-y Creative Commons. (2023). CC BY-NC-SA 4.0 Legal Code | Attribution-NonCommercialShareAlike 4.0 International | Creative Commons . https://creativecommons.org/licenses/by-ncsa/4.0/legalcode Datasette . (n.d.). Datasette: An open source multi-tool for exploring and publishing data. https://datasette.io Davies, T. (2011). Open Data: Infrastructures and ecosystems. ACM Web Science Conference . Davis, K., & Shneyer, E. (2020). Computer science as a tool for developing future civic changemakers. SIGCSE 2020 - Proceedings of the 51st ACM Technical Symposium on Computer Science Education , 1344. https://doi.org/10.1145/3328778.3372636 Enaholo, P. (2017). Beyond mere advocacy: CSOs and the role of intermediaries in Nigeria’s open data ecosystem. The Social Dynamics of Open Data . https://zenodo.org/records/1117797 Enaholo, P., & Dina, D. (2020). Journalists and the intermediation of open data: A Nigerian perspective. In Situating Open Data: Global Trends in Local Contexts (pp. 31–49). African Minds. https://www.scienceopen.com/hosted-document?doi=10.47622/978-1-928502-12-8_2 What is Epicollect5 | Epicollect5 User Guide . (n.d.). What is Epicollect5 | Epicollect5 User Guide. https://docs.epicollect.net Falk, J. (2022). How Game Jams and Hackathons Accelerate Design Processes (arXiv:2205.04966). arXiv. https://doi.org/10.48550/arXiv.2205.04966 GO FAIR initiative. (n.d.). FAIR principles - GO FAIR. GO FAIR. https://www.go-fair.org/fairprinciples/ Goodchild, M.F. Citizens as sensors: the world of volunteered geography. (2007). GeoJournal 69, 211–221. https://doi.org/10.1007/s10708-007-9111-y González-Zapata, F., & Heeks, R. (2015, May). Understanding multiple roles of intermediaries in Open Government Data. Proceedings of the 13th International Conference on Social Implications of Computers in Developing Countries . Gray, J., Bounegru, L., & Chambers, L. (2012). The Data Journalism Handbook (1st ed.). O’Reilly Media, Inc. Herrera-Murillo, D. J., Aziz, A., Nogueras-Iso, J., & Lopez-Pellicer, F. J. Analysing user involvement in open government data initiatives. (2022). in Proc. Int. Conf. Theory Pract. Digit. Libraries, pp. 175–186. https://doi.org/10.1007/978-3-031-16802-4_14 Herrera-Murillo, D. J., Ochoa-Ortiz, H., Ahmed, U., López-Pellicer, F. J., Re, B., Polini, A., and Nogueras-Iso, J. (2024). Process Analysis in Humanitarian Voluntary Geographic Information: the case of the HOT Tasking Manager, AGILE GIScience Ser. , 5, 5. https://doi.org/10.5194/agile-giss5-5-2024 Ho, C., Liu, X., Qiu, Y., and Yang, S. (2024). Research on Innovative Applications and Impacts of Using Generative AI for User Interface Design in Programming Courses. In Proceedings of the 2024 International Conference on Information Technology, Data Science, and Optimization (IDO '24). Association for Computing Machinery, New York, NY, USA, 68–72. https://doi.org/10.1145/3658549.3658566 Implementing regulation - 2023/138 - EN - EUR-Lex . (n.d.). EUR-Lex — Access to European Union law — choose your language. https://eur-lex.europa.eu/legalcontent/EN/TXT/?uri=uriserv:OJ.L_.2023.019.01.0043.01.ENG D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 70 ISO 9241-210:2010, (2010). Ergonomics of human-system interaction - Part 210: Human-centred design for interactive systems. Standard. International Organization for Standardization. Geneva, CH. https://www.iso.org/standard/77520.html Hou, Y., & Wang, D. (2017). Hacking with NPOs: Collaborative Analytics and Broker Roles in Civic Data Hackathons. Proceedings of the ACM on Human-Computer Interaction , 1 (CSCW), 53:153:16. https://doi.org/10.1145/3134688 Jaskiewicz, T., Mulder, I., Morelli, N., & Pedersen, J. (2019). Hacking the hackathon format to empower citizens in outsmarting “smart” cities. Interaction Design and Architecture(s) , 43 , 8–29. https://doi.org/10.55612/s-5002-043-001 Kubler, S., Robert, J., Neumaier, S., Umbrich, J., & Le Traon, Y. (2018). Comparison of metadata quality in open data portals using the Analytic Hierarchy Process. Government Information Quarterly , 35(1), 13 - 29. https://doi.org/10.1016/j.giq.2017.11.003 Kitsios, F., & Kamariotou, M. (2018). Open data hackathons: An innovative strategy to enhance entrepreneurial intention. International Journal of Innovation Science , 10 (4), 519–538. https://doi.org/10.1108/IJIS-06-2017-0055 Li, J., Cao, H., Lin, L., Hou, Y., Zhu, R., & El Ali, A. (2023). User Experience Design Professionals’ Perceptions of Generative Artificial Intelligence. Proceedings of the CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3613904.3642114 Lodato, T. J., & DiSalvo, C. (2016). Issue-oriented hackathons as material participation. New Media & Society , 18 (4), 539–557. https://doi.org/10.1177/1461444816629467 Meijer, A., & Potjer, S. (2018). Citizen-generated open data: An explorative analysis of 25 cases. Government Information Quarterly , 35 (4), 613–621. https://doi.org/10.1016/j.giq.2018.10.004 Linked Open Vocabularies (LOV). (n.d.). Linked Open Vocabularies. https://lov.linkeddata.es/dataset/lov/ Loukis, E. & Saxena, Stuti & Rizun, Nina & Maratsi, Maria & Ali, Mohsan & Alexopoulos, Charalampos. (2023). ChatGPT Application vis-a-vis Open Government Data (OGD): Capabilities, Public Values, Issues and a Research Agenda. 10.1007/978-3-031-41138-0_7. Loutsaris, Michalis Avgerinos & Alexopoulos, Charalampos & Maratsi, Maria & Charalabidis, Yannis. (2023). Semantic Interoperability for Legal Information: Mapping the European Legislation Identifier (ELI) and Akoma Ntoso (AKN) Ontologies. 41-53. 10.1145/3614321.3614327. Maratsi, Maria & Ahmed, Umair & Alexopoulos, Charalampos & Charalabidis, Yannis & Polini, Andrea. (2024 b). Towards Cross-Domain Linking of Data: A Semantic Mapping of Cultural Heritage Ontologies. 165-176. 10.1145/3657054.3657077 Maratsi, Maria & Alexopoulos, Charalampos & Charalabidis, Yannis. (2024 d). A Structured Analysis of Domain-Specific Linked Open Vocabularies (LOV): Indicators for Interoperability and Reusability. 10.1007/978-3-031-56478-9_10. Maratsi, Maria, Alexopoulos, Charalampos & Charalabidis, Yannis. (2024 a). On the Semantic Analysis of Open (Government) Data Portals’ Metadata Provision and Schema. (to be published in a few weeks in the ICEGOV’24 ACM proceedings) Maratsi, Maria, Gialoussi, Nina, Alexopoulos, Charalampos & Charalabidis, Yannis. (2024 c). A Proposed Methodology for Sub-Ontology Development in Comprehensive Scientific Investigation Methods and Tooling (to be published in a few weeks in the MTSR’24 Springer proceedings) Maratsi, Maria & Lachana, Zoi & Alexopoulos, Charalampos & Charalabidis, Yannis. (2023). A functional and semantic analysis of artifact representation schemata in folklore museum websites. International Journal of Metadata, Semantics and Ontologies. 16. 105-117. 10.1504/IJMSO.2023.135339. Marr, B. (2024). Generative AI in Practice: 100+ Amazing Ways Generative Artificial Intelligence is Changing Business and Society. John Wiley & Sons. Meyer, P. (2002). Precision Journalism: A Reporter’s Introduction to Social Science Methods (4th ed.). Rowman & Littlefield Publishers. D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 71 Navalkha, C. (2021). Data for Housing Justice: Examining Activists’ Use of Open Government Data for Housing Justice in Boston, MA and New York, NY [Thesis, Massachusetts Institute of Technology]. https://dspace.mit.edu/handle/1721.1/138973 ODK - Collect data anywhere . (n.d.). ODK - Collect data anywhere. https://getodk.org OpenstreetMap Foundation. (2024). Corporate Members—OpenStreetMap Foundation . https://osmfoundation.org/wiki/Corporate_Members OpenStreetMap Wiki. (2024). Organised Editing/Activities—OpenStreetMap Wiki . https://wiki.openstreetmap.org/w/index.php?title=Organised_Editing/Activities&oldid=2656488 Palova, D., & Vejacka, M. (2022). Experience with Open Data in Project Based Learning. 2022 45th Jubilee International Convention on Information, Communication and Electronic Technology, MIPRO 2022 - Proceedings , 689–694. https://doi.org/10.23919/MIPRO55190.2022.9803595 Papageorgiou, G., Euripides, L., Magnussen, R., & Charalabidis, Y. (2023). Open data journalism: A domain mapping review. Proceedings of the 16th International Conference on Theory and Practice of Electronic Governance , 159–166. https://doi.org/10.1145/3614321.3614340 Park, C. H., Longo, J., & Johnston, E. W. (2020). Exploring Non-State Stakeholder and Community-Led Open Governance: Beyond the Three Pillars of Open Government. Public Performance & Management Review , 43 (3), 587–612. https://doi.org/10.1080/15309576.2019.1677253 Pellegrino, M., & Antelmi, A. (2023). At School of Open Data: A Literature Review: Proceedings of the 15th International Conference on Computer Supported Education , 172–183. https://doi.org/10.5220/0011747500003470 Pence, H. E., Williams, A. J., & Belford, R. E. (2015). New Tools and Challenges for Chemical Education: Mobile Learning, Augmented Reality, and Distributed Cognition in the Dawn of the Social and Semantic Web. In Chemistry Education: Best Practices, Opportunities and Trends . https://doi.org/10.1002/9783527679300.ch28 Poikola, A., Kola, P., & Hintikka, K. A. (2010). Public data: An introduction to opening information resources . Ministry of Transport and Communications. Publications Office of the European Union. Open data maturity report 2023. Publications Office, LU, 2023. https://data.europa.eu/sites/default/files/odm2023_report.pdf QField - Efficient field work built for QGIS . (n.d.). QField - Efficient field work built for QGIS. https://qfield.org Ricker, B., Cinnamon, J., & Dierwechter, Y. (2020). When open data and data activism meet: An analysis of civic participation in Cape Town, South Africa. Canadian Geographies / Géographies Canadiennes , 64 (3), 359–373. https://doi.org/10.1111/cag.12608 Rogers, S. (2008, December 18). Turning official figures into understandable graphics, at the press of a button. The Guardian . https://www.theguardian.com/help/insideguardian/2008/dec/18/unemploymentdata Rogers, S. (2010, August 13). Florence Nightingale, datajournalist: Information has always been beautiful. The Guardian . https://www.theguardian.com/news/datablog/2010/aug/13/florencenightingale-graphics Rogers, S. (2011, September 26). The first Guardian data journalism: May 5, 1821 . The Guardian. http://www.theguardian.com/news/datablog/2011/sep/26/data-journalism-guardian Romero, M., Usart, M., & Ott, M. (2015). Can Serious Games Contribute to Developing and Sustaining 21st Century Skills? Games and Culture , 10 (2), 148–177. https://doi.org/10.1177/1555412014548919 Runeson, P., Olsson, T., Linåker, J., (2021), Open Data Ecosystems — An empirical investigation into an emerging industry collaboration concept, Journal of Systems and Software, Volume 182, 111088, ISSN 0164-1212, https://doi.org/10.1016/j.jss.2021.111088. Saddiqa, M., Larsen, B., Magnussen, R., Rasmussen, L. L., & Pedersen, J. M. (2019). Open data visualization in danish schools: A case study. Journal of WSCG , 2019 (WSCG2019CS), 17–26. https://doi.org/10.24132/CSRN.2019.2902.2.3 D4.2 An approach to steer the behaviour of non-government data holders towards open data through a technical strategy 72 Saddiqa, M., Magnussen, R., Larsen, B., & Pedersen, J. M. (2021). Digital innovation in education: Perspectives, opportunities and challenges of educational open data and sensor data. CEUR Workshop Proceedings , 2991 , 74–83. Saddiqa, M., Rasmussen, L., Magnussen, R., Larsen, B., & Pedersen, J. M. (2019). Bringing open data into Danish schools and its potential impact on school pupils. Proceedings of the 15th International Symposium on Open Collaboration, OpenSym 2019 . https://doi.org/10.1145/3306446.3340821 Salamon, L. M., & Anheier, H. K. (1992). In search of the non-profit sector. I: The question of definitions. Voluntas: International Journal of Voluntary and Nonprofit Organizations , 3 (2), 125– 151. https://doi.org/10.1007/BF01397770 Shaharudin, A., van Loenen, B., & Janssen, M. (2023). Towards a Common Definition of Open Data Intermediaries. Digital Government: Research and Practice , 4 (2), 6:1-6:21. https://doi.org/10.1145/3585537 Stige, Å., Zamani, E.D., Mikalef, P., & Zhu, Y. (2023). Artificial intelligence (AI) for user experience (UX) design: a systematic literature review and future research agenda. Information Technology & People. https://doi.org/10.1108/ITP-07-2022-0519 Thematic Annotation of open Government data, GitHub repository created by IAAA lab, University of Zaragoza, Spain, https://github.com/IAAALab/Thematic_Annotation_of_Government_Data, accessed on 26/09/2024 udata . (n.d.). PyPI. https://pypi.org/project/udata/ Vallejo-Figueroa, S., Rodriguez-Artacho, M., Castro-Gil, M., & Cristobal, E. S. (2018). Using text mining and linked open data to assist the mashup of educational resources. 2018 IEEE Global Engineering Education Conference (EDUCON) , 1606–1611. https://doi.org/10.1109/EDUCON.2018.8363427 van Loenen, B., Vancauwenberghe, G., & Crompvoets, J. (Eds.). (2018). Open Data Exposed . T.M.C. Asser Press. https://link.springer.com/book/10.1007/978-94-6265-261-3 van Loenen, B., Zuiderwijk, A. ., Vancauwenberghe, G., Lopez-Pellicer, F. J. ., Mulder, I., Alexopoulos, C. ., Magnussen, R., Saddiqa, M., Dulong de Rosnay, M., Crompvoets, J., Polini, A., Re, B., & Casiano Flores, C. (2021). Towards value-creating and sustainable open data ecosystems: A comparative case study and a research agenda. JeDEM - EJournal of EDemocracy and Open Government, 13(2), 1–27. https://doi.org/10.29379/jedem.v13i2.644 Weisz, J.D., He, J., Muller, M., Hoefer, G., Miles, R., & Geyer, W. (2024). Design Principles for Generative AI Applications. Proceedings of the CHI Conference on Human Factors in Computing Systems. https://doi.org/10.48550/arXiv.2301.05578 Wolff, A., Cavero Montaner, J. J., & Kortuem, G. (2016). Urban Data in the primary classroom: Bringing data literacy to the UK curriculum. The Journal of Community Informatics , 12 (3). https://doi.org/10.15353/joci.v12i3.3278 Zaman, H. B., Baharin, H., & Ahmad, A. (2021). Fusion Technology and Visualisation to Share STEM Data Using PETS Robots (i-COMEL) for Open Data Readiness Amongst Primary School Children. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics): Vol. 13051 LNCS . https://doi.org/10.1007/9783-030-90235-3_16