Open Data Ecosystem

×

Data Provision

Key elements identified
1. Inclusive
  1. Comprehensive dataset coverage: Availability of a wide variety of thematic datasets such as from different domains such as from science and transport and so on
  2. Facilitating multi-format data access for different users: Providing datasets in multiple formats enhances accessibility for NGOs, businesses, and researchers with distinct technical skills.
2. Circular
  1. Open Data accessibility and interoperability: The Open Data Platform (ODP) requires backing for data integration and interoperability from a range of stakeholders, including governmental and non-governmental entities.
  2. Scalability and performance of the portal: the open data volume, variety and velocity increases by the time, so open data portals should be scalable, and performance should be high to accommodate these variations over the time.
  3. Comprehensive dataset coverage:Availability of a wide variety of thematic datasets such as from different domains such as from science and transport and so on
  4. Enhanced data creation process (for NGD holders and beyond) through better quality, integration and validation tools.
3. user-driven
  1. Technical openness of the datasets: By opening the datasets, it technically means to follow the dataset standards, formats, and other protocols to enhance the data integration across different systems and to facilitate the better accessibility of the open data. Technical openness contributes to the technical interoperability of open data.
  2. Allowing the user to upload their dataset: A user can upload an updated version of the existing dataset. In this way, other users can use cleaned, modified, augmented, or transformed datasets for specific tasks. This feature also contributes to the OD interoperability.
  3. Better interfaces and feedback mechanisms for open data provision: "Important to design and develop the necessary technical mechanisms including new interfaces and user interaction methods that enhance the way NGD holders are creating and sharing open data that is available, usable and valuable to a wide range of stakeholders, fostering a more inclusive and effective open data ecosystem.
Key Strategies proposed
1. Inclusive
  1. Adopting thematic classification of datasets from EDP or other classification such as SDGs
  2. 1. Enable diverse data formats and API options 2. Build accessible user interfaces to enhance usability for non-technical users.
2. Circular
  1. Employing Open Data Standards and Formats for Ecosystem Circularity: Employing open data standards and formats can streamline data accessibility and sharing within open data ecosystems, enhancing their circularity.
  2. Leveraging Big Data and Cloud Computing for Scalability and Performance: Follow the big data and cloud computing technologies to make the portal more scalable and improve performance.
  3. Adopting thematic classification of datasets from EDP or other classification such as SDGs
  4. Enhance the data creation process by employing several technical mechanisms and technologies to improve critical aspects of data creation, quality and integration, such as augmenting semantic interoperability with the help of Large Language Models to automate or semi-automate processes that improve it, improving metadata annotation in open data portals using AI, or validating/checking technical compliance of data on the Web.
3. user-driven
  1. Adoping standards, requirements, and guidelines to open data technically.
  2. developing a pipeline to engage the other open data providers as well.
  3. Enhance the foster improved OD platforms' interface and feedback mechanisms by adopting user-driven design, inclusion of diverse user groups in society and active feedback loops while making use of new technologies wherever this is possible; e.g., Natural Language Processing, Virtual and Augmented Reality, AI, and Chatbots.

×

Driving Force: Data Quality

Description:

Data quality concerns the data's completeness, accuracy, consistency, timeliness, and availability, and the readily applied FAIR principles (Findable, Accessible, Interoperable, Reusable), but also data privacy and anonymization in the open data ecosystems. High-quality data is essential for creating trust among stakeholders, achieving reliable and effective results and making decisions over this data, which further brings innovation. On the contrary, common barriers related to data quality, such as data not available in machine readable formats or not adhering to metadata standards (e.g., DCAT-AP), problems with incomplete, error-filled, and outdated datasets, and lack of metadata, can ultimately affect trust in open data available. Data quality assurance requires proper validation mechanisms, an effective standardization process per data domain, and maintaining the quality over time. These requirements are related to the ability of open data portals to provide users with functionalities such as improved search and results retrieval, data of improved quality, usage of technological means (such as data analysis tools, semantic technology, standards, linked data, data processing tools, anonymization techniques, etc.) in order to achieve and smoothen data integration, enhance the quality, and boost findability and contextual search (thus enabling semantic interoperability of data).

State of the art:

The Open Data Maturity Report (Publications Office of the European Union., 2021) published in 2021 by Capgemini Invent under the European Commission Initiative measures data maturity levels based on four open data dimensions: Policy, Impact, Portal and Quality. If one is to focus on the fourth dimension, quality, then the emphasis is put on quality matters on data but also its derived metadata. According to the Metadata Quality Assessment (MQA) methodology of the European Data Portal (European Commission, 2023), this is mainly focused on analyzing whether format information is available, whether formats can be recognized (if they belong to internationally accepted lists of formats), and whether formats are machine-processable A data quality assessment could include whether the data complies with the expected format, provide tools for the visualization of data in different formats, and provide tools for data analysis, integration, transformation, data scrubbing, and data fusion (data fusion is the process of integrating multiple data sources to produce more consistent, accurate, and useful information than that provided by any individual data source). In addition, data quality, dataset completeness, accuracy and transparency, data availability, legislation/policy, technical skills, and infrastructure (data exchange between governments and users, software for analysis, and platforms) are also mentioned by Safarov et al., (2017) as conditions for better use and re-use of OGD resources. User privacy and anonymization to protect private or sensitive data is another dimension to be considered.

Beyond the state of the art:

Beyond state-of-the-art approaches for data quality include the involvement of users to improve aspects of quality for available data. Schwoerer (2022) referred to data quality and how data can indeed be usable. Apart from the technical aspects of accessibility and navigation techniques, the ability of users to understand and interpret the information provided is put into focus (Schwoerer, 2022) An open data portal should provide the functionality of allowing the users to improve the quality of data (e.g., enrich metadata) if they discover the dataset lacks it, and there should be guidelines for the user to do it correctly. This way, users can contribute to data quality and improve the status of the datasets they are interested in using, making them more usable and discoverable to other users. In addition, open data portal administrators and data users must have the basic skills for data processing and for checking the quality of data. To achieve this, data quality, validity and completeness of the data may be evaluated with the use of technologies available via the data portal.

Key elements identified
1. user-driven
  1. User needs quality measures and propoer profiling of the data: By providing proper data profile and quality measures of the open data can enhence its useability.
Key Strategies proposed
1. user-driven
  1. Follow some kind of quality standards and data profiling techniques to achieve the user-needs regarding the data quality and usability.

×

Data for Artificial Intelligence

Description:

Artificial intelligence is considered as a driving force to drive the synergies in the field of open data. Its purpose is two-fold, firstly, using AI to improve the open data at various stages such as quality, metadata, alignment with standards, etc. on the other hand use of AI to develop applications and use-cases using the open data to bring innovation and value-generation. AI4Data means improving open data in terms of quality (not limited to only quality, interoperability, integration, metadata generation, recommendation etc.). Data4AI means using open data as a source and bringing AI analytical capabilities to foster innovation. Data4AI covers the aspects of making open data available for AI models training and testing. Advancement in AI4Data and Data4AI can open new opportunities in data-driven solutions or easy data product developments.

State of the art:

Artificial Intelligence has revolutionized how open data is used, whether in data quality enhancements or innovative applications. The AI4Data paradigm focuses on enhancing open data at various stages, such as generating or improving metadata, enhancing data quality, making data interoperable, and standardizing it. Several studies have explored these topics, and they are still being investigated avidly following the advent of more advanced models. For instance, Ahmed et al., (2024) created a hybrid method (BRYT) combining conventional and advanced methodologies such as BERT, RAKE, YAKE, TextRank to extract keywords for open data sets. Song et al., (2024) utilize contemporary LLMs such as LLAMA and Mistral to extract metadata from more descriptive fields inside the dataset. Turgunbaev (2024) discusses various conventional machine learning methodologies for the particular usecase of metadata extraction. Giglou et al., (2024) explore state of the art LLMs such as GPT, Llama, Mistral to perform ontology mapping ultimately as a use case for data interoperability. Bakker & Scala, (2024) perform ontology learning from textual data using contemporary LLM: GPT-4o. Alharbi et al., (n.d.) employ LLM- GPT to create an ontological structure – Knowledge Graph from medical texts. Maratsi et al., (2024) explore multiple ontologies and their possible interconnected mapping. Moreover, they employ GPT 3.5 and GPT 4 to generate the mappings and validate its accuracy on manual data. Data4AI paradigm focuses on utilizing open data to enhance the development of AI applications. Open datasets provide significant opportunities for training and improving machine learning models. Access to high-quality open data speeds up AI research by lowering the costs and time required for data collection and preparation. Although it has a lot of room for research, few studies have explored the idea. In one of such studies, Weber et al., (2024) emphasize the significance of open datasets for training machine learning models. They reflect the challenges associated to transparency, data access, and data curation, crucial for advancing open-source AI development. Arbeláez et al., (2024) underscore the significance of employing open data which might be diverse and even unlabeled as a better cost-effective alternative to custom-labelled data which might be expensive and might hinder development. Alam et al., (2023) highlight the role of Open Data and Explainable AI in enhancing transparency, accountability, and interpretability in healthcare AI models. Employing diverse open data, the models can achieve better reliability and accuracy, promoting trust and openness. State-of-the-art reflects how the synergy between AI4Data and Data4AI can create robust feedback loops. Improved data quality and metadata quality (AI4Data) lead to greater model performance, while AI applications and innovations (Data4AI) can further enhance data collection methods through the use of tools using AI. This dual advancement opens opportunities for new data-driven solutions, ranging from automated decision-making systems to scalable data products that address societal challenges in areas like healthcare, climate change, and smart cities.

Beyond the state of the art:

The future holds significant potential for AI4Data and Data4AI. It would revolutionize data curation and AI model training. The current state particularly focuses on improving data quality, interoperability, and accessibility. Moving forward, we can also focus on self-optimizing dynamic ecosystems that use AI and reinforcement learning to maintain the integrity of data, metadata, and interoperability standards. From conventional static systems, it would move to more dynamic systems that would be self-sufficient in addressing changing contexts and user needs. For instance, whatever data-related needs are communicated to the systems via LLMs, which understand the context of user needs, it would trigger the pertinent sub-model to address it and resolve it. Moreover, data in ontological structure and knowledge graphs could prove to be a game changer for transparency, interoperability, and, ultimately, for training models. We could leverage the semantic capabilities of Large Language Models to build domain-specific ontologies to structure open data. It would highly enhance the interoperability and openness of the models that will be trained through those. The domain-specific ontologies and structure would highly enhance the AI model training for any particular domain task. Apart from being accurate it would turn out to be highly transparent and explainable in terms of its predictions. In the future, we may also see the emergence of AI-driven data marketplaces that facilitate the dynamic exchange and monetization of open data through blockchain-based platforms. These platforms would potentially enable data sharing by ensuring provenance, secure access, and fair compensation. This approach would significantly enhance data-driven innovation and promote economic growth on a global scale.

×

Producer's stakeholders Groups

Description:

this group of stakeholders are responsible for producing open data. This can be governments or businesses which usually contribute to the open data ecosystems by providing open data. Their role is to make sure that the data being produced is relevant, useful, accurate, and accessible to the other stakeholders. Producers. They require to maintain data quality and adherence to data maintenance aspects all along the way of data life cycle.

State of the art:

Data producers have a first-level position in hierarchy in the open data ecosystem as shown in Figure 6. They are also known as data providers (Heijlen & Crompvoets, 2021) or data publishers (Charalabidis et al., n.d.; Hrustek et al., 2023). They are responsible for the open data supply in the open data ecosystem, and this open data supply is considered their best effort contributed to the ecosystem. They are responsible for dataset creation or generation (Buteau et al., 2018; Oliveira et al., 2017) and/or publishing (Corbett et al., 2020) open data to fulfill end-user demands in the open data ecosystem. They may share open data with intermediaries to enrich the data and prepare it (Alexopoulos et al., 2014) without incurring any costs, licenses, or restrictions on its use(Kamariotou & Kitsios, 2022; Kitsios et al., 2021). There is a scarcity of skills among open data producers required to make use of data themselves (Kamariotou & Kitsios, 2022). However, data provision by them is not based on demand and needs. Data should be considered for publication if it can be used for service development for citizens, provide value to companies, or contribute to economic improvement (Köster & Suárez, 2016). Often, open data producers are required to develop a framework (legal or regulatory) pertaining to data opening, data sharing, and data use (Dawes et al., 2016; Fang et al., 2024). Not all open data producers publish data freely; sometimes, they publish it under a special license to gain benefits and generate revenue (Heimstädt et al., 2014). In literature, governments as data producers, providers, and publishers have been widely discussed, but a few studies also mention businesses or private enterprises as data producers in the open data ecosystem (Ali et al., 2024).

Beyond the state of the art:

Non-governmental actors should also actively participate in open data provision. In the open data ecosystem, the inclusive nature of an open data ecosystem emphasizes data provision from non-governmental sources (Van Loenen et al., 2021). Data integration techniques should be adopted so that new datasets can be generated from existing ones. The provision of high-quality and demand-oriented datasets can contribute to value creation and sustainability in open data ecosystems. Strategies should be developed to create new datasets (both governmental and non-governmental) and foster synergies among stakeholders. Non-governmental organizations should also be included as part of the producers' stakeholder group, rather than relying solely on governments to provide all required datasets. Government-provided datasets sometimes suffer from low quality, limiting their potential for value creation.

×

Intermediaries

Description:

Open data intermediaries provide specialized resources to enhance the supply, flow, and reuse of open data and the connection between various open data stakeholders. Various actors can take up the role of open data intermediaries, including government organizations, businesses, non-profit organizations, media organizations, and research organizations.

State of the art:

Open data intermediaries are often characterized as being the “bridge” between open data providers and end-users by (pre-)processing open data to facilitate its reuse. Nevertheless, in practice, open data intermediaries often offer more diverse kinds of products and services that improve the process of open data provision and re-use, sometimes without directly handling the data themselves. For example, some open data intermediaries offer software that facilitates open data providers to disseminate open data. Some others offer consulting services to open data providers and users. Media organizations have a more niche function; in addition to the traditional ways of providing open data, they act as communicators of open data to the public. As their main goal is to maximize their audience by providing accurate information, they translate open data insights into stories and visualizations that can be understood by the majority of their audience

Beyond the state of the art:

Greater awareness and inspiration on the potential value-added services that can be offered based on/around open data should be promoted across various sectors/domains to encourage the emergence of more open data intermediaries that help address real-world challenges. Some potential collaborations between open data actors may happen with the help of open data intermediaries that bring different actors together or initiate possible connections or engagement. To some extent, federated architecture could address the current shortcomings where open data across different domains and/or jurisdictions are segmented due to multiple open data providers. The federated architecture means that users can easily integrate open data from multiple domains and/or jurisdictions based on interoperable data standards, but the maintenance of the data still falls under the responsibility of the original provider. This would be valuable to address complex global and local challenges requiring a multidisciplinary and multiscale data-based approach. Open data intermediaries can play a role in initiating or maintaining the federated architecture initiative. Certain open data actors may face financial constraints to purchase or subscribe proprietary software to supply, process, or reuse open data. Hence, the availability of open-source software may alleviate such constraints. Open data intermediaries could play a role in developing open-source software. Nevertheless, all these potential contributions of open data intermediaries require sustainable business models, especially if they were to be undertaken by government or non-profit organizations, which are, at the moment, often rely on sponsorship to carry out open data-related activities/initiatives.

Key elements identified
1. Circular
  1. Value (co-)creation by the ODE actors: Understanding potential contributions of open government data users to the open data ecosystem.
  2. Various open data intermediation value propositions addressing various real-world challenges and providing innovative solutions.
  3. Advisory, consultancy or support services to existing or potential open data actors.
  4. Initiating engagement and interaction among open data actors
  5. The availability of open source software for open data supply, processing, and reuse.
  6. Sustainable business models for open data intermediaries.
  7. Fiscal incentives in exchange for contributions to open data initiatives.
2. user-driven
  1. The availability of open data platform based on federated architecture.
  2. Soliciting various needs of existing and potential open data actors.
3. Skill-based
  1. Providing assistance to data intermediaries to improve the quality of data: The use of methods/approaches that increase data quality, such as validation and augmentation, is crucial in order to make open datasets better and to ensure that they are accessible to all different kinds of users simultaneously.
Key Strategies proposed
1. Circular
  1. NGOs as Value-Adding Intermediaries: NGOs’ current and potential contributions can be in the form of different values that constitute social use values, as these organisations cover a wide range of activities and exist as intermediaries rather than in end-user or producer only form. They are helping the users and the data provider by collaborating and addressing both of them and by adding value through these collaborations."
  2. Open data advocates (such as open data NGOs, researchers, and public agencies) could engage with domain-specific industry players (such as mobility, agriculture, energy sectors) to showcase the potential use of open data and examplary value-added services.
  3. Open data intermediaries (within or outside public sector) should consider introducing functions to provide technical and non-technical supports to other open data actors in supplying, processing, and reusing open data.
  4. Beyond holding events such as hackathons or conferences that bring various actors in one place to connect, open data intermediaries may also consider initiating or leading collaborative initiatives that harness the resources and expertise of multiple open data actors.
  5. The development of open source software should be supported, especially by providing funds for open source software development and maintenance.
  6. Incubator or similar programs to support the design and development of open data intermediaries' business models could be initiated to help existing and future open data intermediaries. These programs could be catered not only to for-profit but also non-profit open data intermediaries, as the former also requires sustainable business models.
  7. Governments should consider offering financial incentives to citizens and companies that support open data or open source projects.
2. user-driven
  1. Open data intermediaries could consider providing an open data platform based on federated architecture.
  2. Open data intermediaries could play an active role in identifying and understanding various needs of open data actors. They can then help to address these needs or communicate these needs to those who may be able to address (such as data providers).
3. Skill-based
  1. 1. Create and distribute data augmentation tools adaptable to different dataset types. 2. The tools should capable of supporting a wide variety of data types and intended uses.

×

User Stakeholders

Description:

The end beneficiary of open data is the user’s stakeholder groups. These groups consist of, but are not limited to, researchers, individuals, citizens, policymakers, businesses, and citizens. They utilize open data as the underlying benefit to making informed decisions and bringing innovations. Data quality is also a matter to be considered for the end-user use to achieve the goal. At this stage, proper educational and training material availability and data accessibility tools make their engagement flourish.

State of the art:

A long list of open data users (Governments, business, entrepreneurs, applications, developers, citizens, universities, researchers, civic hackers, civil society, and consumers) in open data ecosystems have been shown in the Figure 6 adopted from the (Ali et al., 2025). They are also referred to as data consumers (Köster & Suárez, 2016; Lee, 2014; Oliveira & Lóscio, 2018; van Schalkwyk et al., 2016), data demanders (Charalabidis et al., n.d.), in the literature. The overall purpose of this user group in the open data ecosystem is to gain access to the data and/or gain the knowledge about the data quality within ODE (Runeson et al., 2021). In ecosystems pertaining to OGD, the OD users actively participate in data downloading, browsing, and analyzing to within an aim to develop services and applications which could be commercial or for public good (Dawes et al., 2016; Fang et al., 2024). Data use requires to be incentivized as mentioned by the study, and sometime data users are overwhelmed with too many OGD initiatives (Kučera, 2017). Data users can also be competitors within a particular domain of OD (Runeson et al., 2021). Data consumers can also be innovators in the open data ecosystem by innovating new applications and services by utilizing open data (Heijlen & Crompvoets, 2021). The data consumers also play a vital role in interpreting data and providing feedback on the datasets (Corbett et al., 2020).

Beyond the state of the art:

One important clarification that has to be made, and addressed in future development of the SOTA model, is that media organizations overlap with other types of stakeholders. Depending on their business model and organizational structure, they can be NGOs, public, or more commonly, private organizations. In the SOTA model, the media organizations are highlighted only for their role as communicators of news and not for the other roles they can have in the open data ecosystem that can overlap with other stakeholders.

×

Data Exchange

Description:

Data exchange methods, protocols, and standards can help in achieving open data sharing and interoperability among open data stakeholders. It can help in boosting the open data interoperability and maximizing its usage across the open data ecosystems. Better data exchange reduces barriers such as noise, duplication, and interoperability and makes possible cross-domain and cross-organization collaboration and supports the development of innovative applications. The challenges related to data exchange could be data in silos, inconsistent use of standards, and privacy and security concerns.

State of the art:

State of the art in the open data exchange, revolves around the FAIR (findability, accessibility, interoperability, and reusability) principles and other open data standards, guidelines, and requirements adoption (Open Data Charter, 2013; Open Government Working Group, 2007). Some of the standards used such as DCAT (and custom metadata standards e.g. DCAT-AP), open API specifications-based interface designing, and other domain specific standards (OGC) widely adopted to achieve the data exchange and interoperability of open data. Linked data principles and technologies (RDF, SPARQL, OWL) helps in data linking and integration across open data amplifying open data machine readability and reuse (Abid & Laouar, 2018; Akatkin & Yasinovskaya, n.d.; Braga & Ribeiro Gouveia, 2022; Fayyaz et al., 2018). Some of the examples and practices in the open data exchange are European open data portal and US open data portal providing a centralized point to the open data access and implementing the data exchange functionalities at national, regional, and even international levels. Open data tools like CKAN, DKAN, Socrata, OpenDataSoft, etc. are provide built in solution/plugins and customization option to achieve the tailored open data exchange (Ali et al., 2022a). Despite all these efforts, many challenges are still required to be resolved, such as data fragmentation issues, data integration, cross-border open data interoperability, diverging quality of data, big data interoperability, data storage, and other semantic and legal and organizational semantic interoperability barriers (Abella et al., 2018; Ali et al., 2022b, 2024; Avgerinos Loutsaris et al., 2023).

Beyond the state of the art:

Dynamic nature Open data exchange requires beyond state-of-the-art intelligent and effective solutions and approaches integrated within current open data spaces. Artificial intelligence in this regard helps to resolve the semantic issues such as AI-driven data harmonization, compatibility checking, automated semantic annotation and metadata standards alignments, data integration, and linking. APIs and middleware solutions implementation are crucial to be integrated within open data portals to adapt to the highly dynamic nature of user needs in real time. Advances in cloud-edge computing can enable data processing and analysis within the portal boundary and beyond by enhancing the interoperability. Modern policies and software-based solutions for visualizing and monitoring data exchange, machine readability monitoring, automated compliance checks tools, interoperability validation tools, open data integration techniques, and integrated predictive analytics, along with co-development and stakeholder involvement, will boost the data exchange and overall interoperability of open data.

×

Value Co-Creation/Impact/Innovation

Description:

Value (co-) creation covers the collaborative and participatory efforts by the involved stakeholders and actors to bring and sustain the innovation, impact, and value through the open data. It focuses on the shared contributions and collective impact of open data-driven initiatives. This might contain the development of tools, applications, and services to address the challenges (societal, economic, insufficient collaboration, limited stakeholder engagement) and create value for all involved actors. Enhancing value (co-) creation, impact, and innovation are essential for boosting the open data ecosystem potential.

State of the art:

Open data co-creation is the process where multiple stakeholders take part, in a horizontal, user-centered way (Ansell & Torfing, 2021; Ochoa-Ortiz, 2023). Traditionally, co-creation has been focused on citizen-government collaboration (McBride et al., 2018, 2019), but recent research emphasizes the need to engage the full quadruple helix —academia, industry, government, and society (Miller et al., 2018). Open Data plays a key role in this process, yet its successful implementation depends on factors such as stakeholder motivation, proper communication, and technical infrastructure (McBride et al., 2019). While theoretical models for co-creation exist, practical implementations remain limited (Toots et al., 2017). Future developments should integrate standardized co-creation mechanisms into existing data platforms and address both technical and societal barriers to foster sustainable, value-creating open data ecosystems (Van Loenen et al., 2021).

Beyond the state of the art:

Future considerations in value co-creation should focus more on the inclusion of expanding stakeholder involvement, such as decentralized organizations and AI systems, which can enhance value co-creation through collaboration and innovation. The development and adoption of advanced co-creation frameworks tailored to diverse motivations, as well as the technical and social capabilities of stakeholder groups, will enable smoother collaboration. Enhancing co-creation practices on a global scale while addressing both technical and societal challenges, along with adopting AI for innovation, will contribute to more resilient and impactful open data ecosystems. Real-time feedback interventions can drive innovation and adaptability, fostering ecosystem dynamics that sustain their relevance and impact.

×

Collaboration and Network Effect

Description:

Collaboration and network effects in open data ecosystems emphasize decentralized peer-to-peer networks in operations and the formation of communities around shared interests (Briscoe and De Wilde, 2006) that come into action when diverse stakeholders want to communicate and generate value from shared data sources, goals and interests, policy- and governance-based coordination, among others. Collaboration and networking better implementation open doors to innovation, better decision-making, data reuse, and value (co-)creation. On the other hand, open data initiatives face challenges regarding lack of collaboration and network effects, which leads to a reduction in open data's envisioned benefits.

State of the art:

Literature has found that effective collaboration in open data ecosystems depends on governance structures, incentive mechanisms, and technological platforms that facilitate trust, standardization, and seamless data exchange (Mo et al., 2025; Runeson et al., 2021). Examples include data commons (Linåker & Runeson, 2022), open data cooperatives (Linåker & Runeson, 2020), federated data-sharing models (Casaletto et al., 2023), and API-driven platforms (Heshmatisafa & Seppänen, 2023), that enhance interoperability and access. However, collaboration in open data ecosystems face barriers such as data silos (Nikiforova et al., 2024) , trust deficits (Linåker & Runeson, 2021), lack of common standards (Sugg, 2022), and proprietary constraints (Beno et al., 2017). Overcoming these challenges requires strong governance, economic incentives (Reggi & Dawes, 2022b), and global-local coordination mechanisms (López-Reyes et al., 2024)s to realize the envisioned benefits of open data fully.

Beyond the state of the art:

While existing research has highlighted the importance of governance structures, incentive mechanisms, and technological platforms (Mo et al., 2025; Runeson et al., 2021), this section extends these insights, offering strategies to further enhance collaboration and maximize network effects within ODEs, addressing some of the gaps identified in recent literature. Current frameworks for governance in ODEs, such as data commons (Linåker & Runeson, 2022) and federated data-sharing models (Casaletto et al., 2023), have proven to enhance data access and interoperability. However, challenges related to unequal power dynamics and resource disparities persist. For example, large corporations and governments often dominate control over data resources, sidelining smaller stakeholders like NGOs and local communities (Micheli et al., 2020). Research on local open government geodata has shown that collaboration across diverse stakeholders, including public authorities, businesses, and local communities, is essential for fostering inclusive decision-making and creating value from shared data (López-Reyes et al., 2024). Governance models that engage a wide range of actors, particularly at the local level, can enhance transparency and trust, thereby strengthening the collaboration necessary for open data ecosystems to thrive. Future ODE designs must integrate multi-stakeholder governance models, such as data cooperatives or public data trusts, which could facilitate equitable participation by ensuring that control over data is distributed more evenly across diverse actors (Micheli et al., 2020). These governance structures can mitigate monopolization risks, supporting collaboration while addressing existing inequalities in data ownership and control. Integrating open technological platforms that allow for seamless data exchange across different stakeholders, enabling innovation, and cross-sector collaboration is key in ODEs (López-Reyes et al., 2024). Generative technologies such as APIs can facilitate the creation of new applications that support diverse use cases. As highlighted by Toorajipour et al. (2024) AIoT systems can significantly enhance the generative potential of ODEs by automating data processing and enabling real-time decision-making. However, while these technologies unlock vast opportunities, they also present technological barriers for smaller actors who lack the financial or technical resources to integrate such tools into their operations (i.e. local governments) (Zuiderwijk et al., 2018). Therefore, ODEs must not only provide the technological infrastructure needed for collaborative innovation but also support capacity-building initiatives to ensure that smaller entities, such as local businesses and NGOs, can engage fully. This can be achieved by offering shared infrastructures, subsidies, and training programs aimed at leveling the playing field and empowering underrepresented groups (Hambleton & Jeyaseelan, 2024). There is also a need for trust-building mechanisms that consider the challenges stakeholders face in open data ecosystems due to lack of transparency and unclear data ownership (López-Reyes et al., 2024). Establishing clear data governance structures that promote collaboration, and accountability is essential to overcoming these trust deficits, especially when local governments and citizens are engaged in data-sharing efforts. Trust remains a foundational element of successful ODEs, especially in light of increasing concerns over data privacy and security (Micheli et al., 2020). Current solutions, including blockchain and privacy-preserving technologies, have the potential to increase trust by ensuring transparency and accountability in data usage. However, concerns over selective openness, where only certain data sets are made accessible, continue to undermine the inclusivity of data-sharing initiatives (Beno et al., 2017). To address this, ODEs should implement robust privacy frameworks alongside clear data governance models that balance privacy with the need for openness. These models should ensure that non-personal data, when aggregated or used in innovative ways, does not inadvertently cause harm or reinforce existing power imbalances (Micheli et al., 2020). Beyond economic incentives, the social value generated by ODEs is increasingly recognized as a key driver of sustainability. This underscores the importance of community-driven value creation (as underlined in ODECO D.5.2), where data is used not only for economic benefits but also to address social issues like urban governance (López-Reyes et al., 2024). Purpose-driven ODEs that focus on social inclusion and public services can contribute to addressing local needs while engaging citizens and fostering democratic participation (Verhulst et al., 2020). The concept of purpose-driven data initiatives is gaining momentum, where data ecosystems are aligned with specific social and environmental goals. These initiatives, however, often face the challenge of balancing competing priorities among stakeholders. As businesses and governments prioritize innovation and profit, civil society and marginalized groups may emphasize transparency and equity (Sorri & Seppänen, 2021). A shift towards purpose-driven ODEs, guided by clear public-interest goals, will help align data practices with societal benefits, such as improving public health, fostering environmental sustainability, and promoting democratic engagement (Arena et al., 2022). Engaging underrepresented communities in decision-making processes through co-ownership or active partnerships will ensure that data initiatives reflect the collective needs of society (as mentioned in ODECO D.5.2). One of the key challenges in sustaining the value of open data over time is the maintenance and curation of data, which requires ongoing investment (Jetzek et al., 2019). ODEs are not only about the continuous generation of data but also about ensuring that data quality is preserved across its use, requiring proper infrastructure and feedback mechanisms to maintain long-term engagement (López-Reyes et al., 2024). Many open data ecosystems rely on voluntary contributions from data providers, but without mechanisms to track data reuse and measure its value, contributors may lose incentives to continue sharing data (Lindman, 2014). To overcome this, ODEs must develop robust feedback systems that help measure the impact and value generation from shared data. Additionally, a balance must be struck between innovative data use and ensuring that data quality is maintained across all sectors (Tseng & Nikiforova, 2025).

Key elements identified
1. Inclusive
  1. Alignment of interests:The farmework must align organizational, societal, and indivitual interests to foster collaboration and shared responsability.
  2. Open Data contribution from non-government actors: The need to escape the narrow limits of only governmental bodies being active contributors to the open data ecosystem; rather, include other stakeholders (non-governmental data groups) as potential contributors.
  3. Interdisciplinary collaboration: Collaboration between domain-experts and data experts. Domain-experts can contribute with contextual knowledge and priorities that need to be addressed with open data. Data experts can contribute their skills for data analysis (cleaning, visualization, etc.).
Key Strategies proposed
1. Inclusive
  1. Shared Goals for Stakeholder Collaboration: Create shared goals and values among diverse stakeholders within an open data ecosystem. This ensures cohesive collaboration and maximizes the ecosystem’s ability to deliver value across its users, contributors, and beneficiaries.
  2. Analyse the barriers, motivations, and technical requirements for non-government data holder groups to share their data and propose technical mechanisms to enable/facilitate this purpose.
  3. Apply strategies to incentivize and facilitate the collaboration of domain experts and data experts.

×

Incentive Arrows

Description:

Incentivizing open data ecosystem mechanisms that can motivate and stimulate the stakeholder’s engagement and contributions. These incentives cover financial support, regulatory mandates, social benefits, and impacts. Incentives encourage stakeholders’ involvement, data sharing, generating value, and bringing innovation in open data ecosystems. The unavailability of incentive mechanisms limits the engagement and utilization of open data and available resources. Developing and aligning incentives are critical for sustaining open data stakeholders’ participation and overall initiatives.

State of the art:

The current incentives for the small sized NPOs (Non-Profit Organizations) to engage and contribute to the open data ecosystem include achieving openness and transparency goals of the organizations and employees (Barnes et al., 2025). Openness is one of the values shared within the open data focused NPOs as part of the cultivated and shared organisational culture of these organisations. NPOs aim to create a social impact as part of the cultivated and shared organisational culture, which means they have an awareness about the social impact of open data sharing. Additionally, the organisational culture of NPOs is often less hierarchical, and employees can propose and take part in projects they enjoy, which means they have a presence of engagement or enjoyment activities to motivate data sharing. Another incentive is the community support and social impact of open data sharing and open data projects. It is motivating for small sized NPOs to continue the project/data sharing, if the impact and support are evident, as their resources are otherwise limited, so they must prioritise which projects to work on/data to open. Thus, community support and influence are an important for

Beyond the state of the art:

For non-profit and non-governmental organizations, important incentives to open data sharing and value generation are training to improve data-literacy skills of the employees together with raising awareness of open data concepts. Smaller sized NPOs and NGOs may not have a dedicated programmer or skilled employee to share the data properly as open or reuse the existing open data to create social value. To motivate these actors, they need to have access to training and the ability to employ skilled employees. Another incentive for NPOs and NGOs is more financial cooperation that would support data-skilled employees and help NPOs run their open data projects long-term. There are existing examples of such cooperation, such as the Open Knowledge Foundation network’s prototype fund, that can help small organisations or activists financially with their projects and data literacy training. Similar solution should be promoted, especially across the countries, to help distribute the resources to the NGOs with less support that needed for open data ecosystem engagement. Moreover, the creation of a common platform for NGOs to share the data, like open data government portals, could work as another incentive. The creation of such a portal could help with the availability of appropriate technical tools and availability of resources that many NGOs lack and, thus, are unable to share their data (Petti, 2024).

Key elements identified
1. Circular
  1. Partnerships should be formed between public administrations / international organisations and other civic open data user groups to co-develop open data initiatives, which can prompt contributions from other actors.
  2. Motivations of non-government actors to become active contributors to the Open Data ecosystem
  3. Enabler: Availability of training in data skills and literacy
  4. Enabler: Availability of appropriate technical tools
  5. Enabler: Alignment of private value and interests with open data sharing.
  6. Enabler: Availability of resources (financial, time, people/workforce)
  7. Enabler: Existence of data-sharing communities.
  8. Enabler: Awareness about the social impact of open data sharing.
  9. Enabler: Presence of engagement or enjoyment activities.
2. user-driven
  1. Partnerships should be formed between public administrations / international organisations and other civic open data user groups to co-develop open data initiatives, which can prompt contributions from other actors.
Key Strategies proposed
1. Circular
  1. Partnerships can encourage non-government data holders to contribute open data, in addition to being users of open data.
  2. Key enablers to contribute data to the ODE were identified, that relate to the motivations and barriers found by users.
  3. By making training in data skills and literacy available, data contribution can be enabled/leveraged.
  4. By making appropriate technical tools available, data contribution can be enabled/leveraged.
  5. By aligning private value and interests with open data sharing, data contribution can be enabled/leveraged
  6. By making resources (financial, time, people/workforce) available, data contribution can be enabled/leveraged.
  7. By making data-sharing communities exist, data contribution can be enabled/leveraged.
  8. By making stakeholders aware about the social impact of open data sharing, data contribution can be enabled/leveraged.
  9. By having engagement or enjoyment activities, data contribution can be enabled/leveraged.
2. user-driven
  1. Willingness on part of public administrations to engage in formal and informal partnerships with other civic actor groups

×

Usage to Provision

Key elements identified
1. Circular
  1. Promoting reusable dataset integration techniques: Techniques like data blending and augmentation improve the ecosystem’s efficiency and user engagement.
  2. Enhanced data creation process (for NGD holders and beyond) through better quality, integration and validation tools.
2. user-driven
  1. User-driven Feedback mechanisms: Feedback mechanisms, including dataser rating systems, dataset comments, and direct user feedback through discussion forum, are essential for enhancing data quality and ensuring user-friendly portals.
  2. User needs quality measures and propoer profiling of the data: By providing proper data profile and quality measures of the open data can enhence its useability.
Key Strategies proposed
1. Circular
  1. 1. Provide training on dataset linking and blending tools 2. Enable version control for community-shared datasets.
  2. Enhance the data creation process by employing several technical mechanisms and technologies to improve critical aspects of data creation, quality and integration, such as augmenting semantic interoperability with the help of Large Language Models to automate or semi-automate processes that improve it, improving metadata annotation in open data portals using AI, or validating/checking technical compliance of data on the Web.
2. user-driven
  1. Integrates user contributions so that it can improve overall dataset relevance and usability. Establish effective feedback mechanisms.
  2. Follow some kind of quality standards and data profiling techniques to achieve the user-needs regarding the data quality and usability.

×

Open Data Ecosystem Drivers

Key elements identified
1. Inclusive
  1. CARE principles account for power dynamics in OD ecosystems: Legal avenues for redistributing value need to recognise and account for the politics of open data production and re-use. Normatively, this can be achieved through an orientation towards ‘data justice’. (Taylor, 2017) In terms of practical implementation, legal frameworks for open data initiatives could refer to CARE principles in addition to the FAIR principles for data quality management. (ODECO, 2023).
2. Circular
  1. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 1/6: Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts.
  2. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 2/6: Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts and these roles influence the types of financial value they can create or capture.
  3. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 3/6: Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts and these roles influence the types of financial value they can create or capture.
  4. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 4/6: Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts and these roles influence the types of financial value they can create or capture.
  5. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 5/6: Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts and these roles influence the types of financial value they can create or capture.
  6. The financial value created or captured by actors depend on the roles they play in the OD Ecosystem, as they can play multiple roles at the same time. 6/6 Actors in an open data ecosystem (e.g., government agencies, companies, non-profit organisations, and citizens) are not wedded to any particular role . In other words, they can play multiple roles at the same time or different roles in different contexts and these roles influence the types of financial value they can create or capture.
  7. In OD Ecosystems, value is seen as multifaceted and dependent on the dynamics of user engagement, circularity, inclusivity, and skill-based development within open data systems: The complexity of ‘valuing’ open data arises from the various stages in the open data usage process, which range from collection and storage to analysis and sharing. Each stage presents unique requirements and risks that need to be managed to maintain the data's utility and integrity while ensuring responsible data practices that consider privacy, equity, and accountability (S. G. Verhulst, 2021). In the open data ecosystems proposal, the concept of value is approached from a broader, ecosystem-oriented perspective rather than merely traditional use-value.
3. user-driven
  1. Enhancing accessibility of metadata: Metadata must remain accessible despite the unavailability of datasets, facilitating the ongoing utilization of historical context for research purposes.
  2. Boundary-making in relation to OD Ecosystems: To define the boundaries of an OD ecosystem, actors should be aware of the socio-technical conditions where their interactions with other actors take place. These conditions refer on one hand to social components such as historical, geographical aspects, social and cultural norms, organization norms and community affiliations, as well as practices, traditions, personal motivations and values. On the other hand, the technical aspects include soft and hard data infrastructures, interoperability practices, standards, laws and regulations. Ecosystem mapping can be done using tools from the discipline of design thinking and theoretical principles from the discipline of information visualization and communication.
Key Strategies proposed
1. Inclusive
  1. In terms of practical implementation, legal frameworks for open data initiatives could refer to CARE principles in addition to the FAIR principles for data quality management. (ODECO, 2023). These CARE Principles do not focus only on inherent qualities of data that enable more sharing and re-use, but also focus on power differentials. As a result, CARE Principles also focus on realising collective benefit from open datasets, as well as ensuring commitment to ethics. (Carroll et al., 2020). There are examples of CARE principles being translated into the design and implementation of open data initiatives - such as the Open Data Mekong project. (Chung & Chung, 2019; DCPC, 2024). 
2. Circular
  1. Commitments to provide continuous and adequate funding for maintaining and developing open data infrastructure and support ought to be institutionalized, for example, through supranational, national, or local laws. Conversely, open data has to be considered as an infrastructure (like roads, clean water, etc.) and be accounted for in the tax policy design that, at the same time, does not disincentivize companies from continue using open data.
  2. Some restructuring, streamlining, and coordination of open data management and provision may be necessary within and across government agencies to save (transaction) costs as open data providers further. At the same time, it may also be worth coordinating with other agencies that are not necessarily open data providers to implement contractual obligations with third-party vendors that collect data through their government-funded projects/undertakings to give such data to the government free of charge (including augmented data on top of initially government open data), so that (some of) the data can then be provided as open data. This saves the government from (re)collecting the data.
  3. One possible way to recuperate some of the costs of providing open data is for government agencies to also provide value-added services based on open data at fees. This involves revisiting, clarifying, and potentially amending existing laws related to government-market roles and competition.
  4. Government agencies should be obligated to use open data whenever it is available and to inform open data providers of any errors in the data. This saves costs for them as open data users as well as to open data providers in ensuring data quality.
  5. Tax incentives may be worth considering to entice companies and citizens to contribute to open data initiatives as funders or data providers. The case of the Wikimedia Foundation and OpenStreetMap Foundation being granted tax-deductible status in some countries, where companies and citizens who donate to these organizations can claim tax deductions, is a good example to consider for other open data initiatives.
  6. Governments should recognize that they are no longer the only open data providers. Hence, they should consider supporting citizen-generated or crowdsourced open data projects, such as OpenStreetMap, by financially supporting or contributing their own data to these projects. At the same time, governments can also leverage open data from these projects as one of their data sources. This may lead to more cost-sharing in open data initiatives within the ecosystem.
  7. A purpose-driven approach, rather than a user-driven one, can enhance open data ecosystems to better reflect the understanding of social value. Purpose-directed models establish clear goals, usually concentrating on encouraging responsible data usage while advancing societal benefits, such as improved healthcare and environmental sustainability.
3. user-driven
  1. Ensures that datasets remain significant and valuable throughout time, hence contributing to the sustainability of the ecosystem. 1. Implement metadata repositories 2. Provide documentation for metadata usability
  2. Awareness is needed on the socio-technical conditions in which interactions between actors take place.

Understanding the key drivers that propel the open data ecosystem forward.

User-Driven

Most existing OD systems are neither user-driven nor balance demand-supply matching. Although users' role is assumed to be critical for the success of OD ecosystems, suppliers drive current ecosys-tems (Gascó-Hernández et al., 2018; Sieber & Johnson, 2015; Van Loenen, 2018; Crusoe et al., 2019). This means that provisioning data from governments to end users is often made through a one-way data portal or platform (Evans & Campos, 2013) without differentiating between various types of users (including cross-border users) or providing data in the way that the data supplier perceives as satisfying (see European Commission, 2020c). Even more, the published data may have no value or are not directly usable (Rhind, 2014). OD providers do not close (enable) the feedback loop – a mech-anism for demand-supply matching – and hence are unaware of the actual use and users of their data, and what data and in what form the users need the data to create value (Harrison et al., 2012; Susha, 2015). That is, they only know that opening datasets increases “usage numbers”, i.e. number of downloads (van Loenen et al, 2018).

Circular

Existing OD systems are often linear. In linear OD ecosystems, OD users typically capture value without adding value to the system themselves. At the moment, only very few stakeholders in the ecosystem profit from OD, and usually, the ones who profit are not the ones who invested most effort, time, and money (Pollock, 2011; Welle Donker & Van Loenen, 2018). As a result of this unfair distribution of value in the OD ecosystem, stakeholders may be reluctant to participate. An OD eco-system is circular (BSI Group, no year) when data is opened by data providers and then processed so that intermediaries, value-adding resellers, and enrichers can keep processed data in use for as long as possible. Users extract the maximum value from derived data/services. Finally, closing the cycle, stakeholders provide additional value to the ecosystem (Alexopoulos et al., 2014). For exam-ple, they can deliver value-added products based on the OD or evaluate data usage processes with data providers (Charalabidis, 2016). The necessary type of ecosystem is circular in nature, building upon a complex network of value generated by different participants that are creating valuable in-formation and products and services. However, note that a circle that is too small can harm circular OD systems because prosumers are then focused on serving data to themselves.

Inclusive

Many current OD systems are exclusive. Research and practice mainly focus on a minimal selec-tion of OD providers and user groups, namely governments as data providers and companies as OD re-users. At the same time, many more stakeholders exist (Van Loenen et al., 2018). Consequently, government data and non-government data (e.g., from businesses, citizens, researchers, and others) are often not integrated and do not cross borders. Moreover, not all stakeholders (e.g., data provid-ers, intermediaries, value-adding resellers, enrichers, facilitators, end-users) from any nature (com-mercial, government, scientific, citizen) participate in OD ecosystem processes (eu2020.de, 2020; Eu-ropean Commission 2020).

Skill-based

Existing OD systems usually depend on best effort. There is a lack of skilled people to use OD. Re-search indicates that many businesses and governments do not have the knowledge and skills to decide on and implement effective strategies of making the best use of their data (Barbero et al., 2018; Davies et al., 2019). This is already reflected in the digital gap within the EU public administra-tions and the significant variations in the quality of digital public services and which is heavily in-fluenced by the discrepancies in digital skills and competences (Chinn et al., 2020; European Union, 2020). There are currently not enough people who have the right data skills (European Commission, 2020b; IDC & Open Evidence, 2017). This problem is likely to get even worse in the future (Deloitte, 2017). Despite the relevance of soft skills (Laker and Powell, 2011) when considering technological changes (Snape, 2017), the skills required are a higher education degree in economics, mathematics, physics, or other relevant scientific disciplines, plus familiarity with the industry concerned (Econ-omist Intelligence Unit, 2012). The combination of all these skills is difficult to obtain (Deloitte, 2017).

×

Recommendation

Technical Recommendations

Improve discoverability by adopting Thematic Structured Classification Models
Standardizing the classification of datasets ensures consistency and improves data discoverability. Adopting established classification frameworks, such as those used by the European Data Portal, facilitates interoperability and alignment with broader data ecosystems. Thematic categorization enhances the usability of datasets, enabling more efficient search, retrieval, and analysis by end users. Moreover, enhancing metadata quality for improved data discoverability to maximize the impact of open data, policymakers should enforce rigorous metadata requirements, ensuring datasets include comprehensive descriptions, provenance, and structured classifications. This will improve data discoverability, support machine-readable formats, and enable more effective data utilization by researchers, businesses, and the public sector.
Promoting standardization through Linked Open Data for Semantic Interoperability
Governments should prioritize the adoption of Linked Open Data principles to strengthen semantic interoperability. By leveraging ontologies such as SKOS and Wikidata, data portals can interlink datasets more effectively, allowing for better contextualization, cross-domain analysis, and enhanced knowledge extraction. Governments and organizations should establish mandatory compliance with widely accepted interoperability standards such as DCAT, FOAF, and Schema.org. Ensuring adherence to these standards will facilitate seamless data exchange across platforms, enhance data usability, and promote a more integrated open data ecosystem. This will strengthen open data Interoperability Through Standardization.
Establishing a Governance Framework for Open Data Validation
A dedicated governance framework should be implemented to oversee the technical validation of open data, ensuring data consistency, reliability, and compliance with best practices. This framework should include automated validation tools, periodic audits, and collaborative stakeholder engagement to maintain high data quality standards.
Establishing Effective Feedback Mechanisms for Data Ecosystem Improvement
To enhance user engagement and data quality, policymakers should mandate the establishment of feedback mechanisms such as user surveys, dataset ratings, and discussion forums on open data platforms. These tools allow users to report issues, suggest improvements, and engage in continuous dialogue with data providers. Integrating AI-driven sentiment analysis and real-time dashboards can help improve platform responsiveness and trust, creating a dynamic and user-responsive ecosystem that fosters continuous data quality improvement and broader adoption.
Standardizing Open Data Portal Interfaces to Simplify Data Publication
Governments should adopt standardized interfaces for open data portals that streamline the data publication process. A uniform framework for data entry, metadata documentation, and access management will reduce inconsistencies and ensure datasets are discoverable and interoperable across platforms. Automation tools like data validation and version control can enhance data quality, while open APIs and export functionalities will maximize data sharing and integration, ultimately improving accessibility and encouraging broader participation in the open data ecosystem.
Improve technical usability and openness
Enhancing data usability requires offering diverse access points, including various data formats, APIs, and user-friendly interfaces that accommodate both technical and non-technical users. Applying FAIR (Findable, Accessible, Interoperable, and Reusable) principles, open standards, and data quality profiling techniques ensures greater accessibility and engagement. By enabling multiple entry points, users can retrieve and interact with data more efficiently, fostering broader adoption and utilization of open data. Policymakers should encourage the adoption of technical openness in data publication by adhering to machine-readable formats, following Open API specifications, and utilizing technical standards that improve data integration across systems. Offering tools such as API explorers and fostering collaboration on technical solutions will ensure that datasets are accessible and reusable. A methodology to quantify the technical openness of datasets should be developed to ensure that open data platforms meet the highest standards of data interoperability. Ensuring inclusivity in open data platforms involves designing intuitive interfaces, offering interactive visualizations, and providing multilingual support. Features such as screen reader compatibility, high-contrast visuals, and accessible navigation cater to diverse user needs, including those with disabilities. By reducing usability barriers, open data becomes more accessible to a wider audience, driving engagement and effective utilization.
Facilitate Data Sharing Among Non-Government Data Holders
Encouraging private entities, research institutions, and civil society organizations to share data requires addressing legal, technical, and organizational barriers. Establishing clear licensing frameworks, anonymization techniques, and secure data-sharing mechanisms can help build trust among non-government data holders. Open-source platforms can further support seamless collaboration by ensuring interoperability and compliance with existing data standards. Partnerships and coordination efforts should also be strengthened to promote contributions from diverse stakeholders.

Governance Recommendation

Build Shared Goals and Values by promoting shared culture
Fostering collaboration among government agencies, private sector actors, researchers, and the public necessitates the development of shared objectives and values. Organizing stakeholder consultations, workshops, and collaborative forums can help align goals and create a common vision for open data initiatives. Promoting a culture of transparency, innovation, and community-driven solutions strengthens long-term commitment to open data practices.
Scope Local Problems and Build Collaborative Networks with local stakeholders
Addressing local challenges through open data requires conducting needs assessments and actively engaging with local stakeholders. By identifying specific problems and collaborating with communities, policymakers can develop targeted data-driven interventions. Building sustainable data-sharing networks and fostering partnerships with local organizations help ensure that open data initiatives effectively respond to regional priorities and constraints.
Facilitate Collaboration Between Domain Experts and Data Experts
Bridging the gap between subject matter experts and data scientists enhances the quality and applicability of open data solutions. Structured initiatives such as hackathons, interdisciplinary workshops, and collaborative research projects provide opportunities for knowledge exchange. Additionally, offering incentives like grants and training programs encourages long-term engagement and fosters innovation in data-driven decision-making.
Empowering Marginalized Stakeholders for Inclusive Open Data Ecosystems
To create a more inclusive open data ecosystem, governments should ensure that marginalized groups, such as small-scale data contributors and NGOs, have the resources and opportunities to engage meaningfully in data-sharing activities. This can be achieved by creating accessible, multilingual feedback channels, offering capacity-building programs, and incentivizing participation through training and support. By adopting collaborative governance models and fostering community-driven data stewardship, policymakers can bridge power imbalances and ensure that the open data ecosystem is equitable and sustainable.
Enabling Data Analysis Through Low-Code Tools and Training
To enhance the accessibility of open data for non-technical users, governments should support the development and distribution of low-code data analysis tools. These platforms will allow users with varying levels of expertise to analyze open data and generate insights. Policymakers should also provide training for educators and intermediaries on how to use these tools effectively, ensuring that a wide range of stakeholders can participate in the data-driven decision-making process. Moreover, Governments should encourage the development and distribution of data augmentation and validation tools that support various dataset types. These tools, developed by intermediaries within the open data ecosystem, should be adaptable to different data types and intended uses. By enhancing data quality through these tools, governments can ensure that open data remains accurate, relevant, and usable, further improving its value to stakeholders across sectors.

×

Resources

Description:

Resources in open data ecosystems perspective covers dimensions like datasets, metadata models, tools, and funding (Cf. par. 6.3.9) to manage the data, and analytical models which can support data accessibility and usability across wider domains. These resources are important to attain innovation and benefit out of the open data (e.g., value-creation). Better resource management, allocation, and utilization are required to ensure transparency and accountability of the open data ecosystem for fair value creation and distribution. There is also a need for funding models to cover public-private partnerships and open innovations.

State of the art :

According to the Open Data Barometer , open data initiatives that have support from the government and sustainable availability and access to required resources have a high chance of achieving their defined goals and impact. Open data initiatives development and implementation needs continuous availability of resources (e.g., finance) (Donker & van Loenen, 2017) . Data curation and sharing solutions are considered important and they provide solution for “a common metadata tracking framework, providing tools and resources to create and manage large, heterogeneous data sets in a coherent manner, and allowing users of (parts of) data sets to ‘connect the metadata dots”(Sansone et al., 2012). The data resources and tools is considered a characteristic of an open data ecosystems (Zuiderwijk et al., 2014). The data resources cover several aspects, including address licensing, linked data, data packaging, patching, pull requests, and data merging options. Sometime resources are costly and face certain risks. The data resources should also include ICT preparedness, ICT usage, architecture, and resource allocation (Zuiderwijk et al., 2014). Open data ecosystems have wicked problems with commercial value creation. A study by Zuiderwijk et al. (2016) mentions that OD policies should be oriented and pay attention to the availability and quality of five types of resources, such as open data, open information technology, internal IT, required knowledge, and governance. Resources on the open data producer's side are needed, but some resources are also needed on the data user's side, such as open datasets, the skills required to analyze the data, stakeholder collaboration and network effects, and competitive advantage based on open data. Open data resources required to create a competitive advantage have also been discussed in the study by Zuiderwijk et al. (2016). Even companies can acquire the required resources from their allies to generate commercial value in open data ecosystems. Resources from different companies may lead to value creation and give companies an opportunity to rely on specific resources that are important to them. The need for resources also varies from company to company. Resource needs are also expandable for all stakeholders to some extent.

Beyond the state of the art:

The availability of resources is mostly focused on the producer's side, and less attention is given to the OD users and intermediaries' side. There is also a need for funding models to cover public-private partnerships and open innovations. It’s also crucial that sustainable funding is provided for other parts of the ODE, including data infrastructure, governance, and ongoing community engagement. Some studies have already considered open data as a resource for creating value in open data ecosystems, and related concepts such as data models, data sources, and data quality are also important in the value creation of the open data ecosystem. We have seen studies that discuss data packaging and patching, but cross-platform interoperability remains an issue. Work is required toward the development of a standardized resource framework covering different kinds of resources (technical, social, financial, and even human). Considering social capital as a quantifiable resource in ODEs is also important.

Key elements identified
1. Circular
  1. Open Data accessibility and interoperability: The Open Data Platform (ODP) requires backing for data integration and interoperability from a range of stakeholders, including governmental and non-governmental entities.
  2. Scalability and performance of the portal: the open data volume, variety and velocity increases by the time, so open data portals should be scalable, and performance should be high to accommodate these variations over the time.
Key Strategies proposed
1. Circular
  1. Employing Open Data Standards and Formats for Ecosystem Circularity: Employing open data standards and formats can streamline data accessibility and sharing within open data ecosystems, enhancing their circularity.
  2. Leveraging Big Data and Cloud Computing for Scalability and Performance: Follow the big data and cloud computing technologies to make the portal more scalable and improve performance.

×

Enabling Environment

Description:

Enabling the environment is meant to cover aspects of sustainable open data ecosystem such as open data market (e.g., business models for sustainability), culture, regulations, and other capacity building, and community participation activities that support or hinder the success of open data initiatives. Community participation is important as the open data ecosystems assume the utilization of data outside the systems that generated them to create value. The proper arrangements of the enabling requirements reduce barriers to open data sharing and use, promote FAIR principles, drive innovation, and foster transparency across diverse cultural contexts. In the enabling environment there is relevant to consider the ethical considerations such as bias, privacy concerns, and misuse of open data (Zuiderwijk & Janssen, 2014).

State of the art:

An open data ecosystem is not a neutral goal but lends itself to multiple interpretations regarding knowledge production, political economy, and the distribution of benefits. It is, therefore, important to understand that any enablement requirements are part of a bigger system of values in which the definition of transparency and participation is negotiated. Knowledge production is political and shaped by power dynamics that enable or limit participation in open data initiatives (Bates, 2013; Johnson, 2014b; Kitchin, 2024) . More specifically, some actors lack the skills to make sense of open data (Perovich et al., 2020). This is particularly true for marginalized communities who cannot actively participate in open data initiatives or who might not be represented by data due to data gaps (Milan, 2020; Ruijer et al., 2024). Open data governance is opaque, with accountability dispersed across private-public partnerships (J. Taylor, 2017; L. Taylor et al., 2022). This lack of transparency makes it difficult to track data movement, integrity, and security, leading to risks such as unintended data repurposing (e.g., U.S. open science data used for facial recognition in China) (L. Taylor et al., 2022). Further, AI and data infrastructures complicate and dilute the process of tracing data origins, making value assessment challenging (Publications Office of the European Union., 2024). Power imbalances shape data access and usage, leading to selective openness, where dominant actors (e.g., corporations) benefit more than smaller entities or individuals (Broomfield, 2023; Gurstein, 2011).

Beyond the state of the art:

Starting from the consideration that open data ecosystems also bring conflicts among values, different governance strategies can be envisaged in terms of the study and practice of open data ecosystems: Study the extent to which open data regulations mandate or promote data sharing that is not meant to be repurposed in the data economy. For instance, the implementing regulation on High-Value Datasets (2023/138) prioritizes economic value, thus enabling greater economic/market value rather than other values, such as participation (Broomfield, 2023). Understand the role of commercial actors in open data by adding obligations for open data sharing in the public interest, as seen in Barcelona’s ‘data sovereignty clauses,’ which require private service providers to share data in an open, machine-readable format within procurement contracts (Monge et al., 2022). Rely on private law mechanisms to promote equitable data use, as in the case of stronger share-alike obligations that prevent the risk of open data exploitation (Data Science Law Lab, 2024).

Key elements identified
1. Inclusive
  1. Open data event design: Open data events (such as hackathons, mapathons, etc.) can contribute to increase the inclusiveness of OD ecosystems if designed appropriately (group formation methods, agenda, activities, etc.)
  2. Institutionalization of open data events: Public administration bodies can organize and financially support open data events, which provide networking and learning opportunities, especially for non-specialist users.
  3. Legal avenues for redistributing value in the open data ecosystem need to account for the current inequities in open data initiatives. 1/3: Redistributing value in the open data ecosystem must address existing inequities, as open datasets are inherently political and shaped by systemic biases. While principles like FAIR improve data quality and representativeness, they often treat datasets as neutral, ignoring critical gaps. Missing data, such as underrepresentation of Global South countries during the Covid-19 pandemic or indigenous communities in the Global North, skews decision-making and global policy. Similarly, localized examples, like Brussels' daycare data failing to integrate mobility information, highlight the practical limitations of open datasets in addressing accessibility and equity issues.
  4. Legal avenues for redistributing value in the open data ecosystem need to account for the current inequities in open data initiatives. 2/3: Redistributing value in the open data ecosystem must address existing inequities, as open datasets are inherently political and shaped by systemic biases. While principles like FAIR improve data quality and representativeness, they often treat datasets as neutral, ignoring critical gaps. Missing data, such as underrepresentation of Global South countries during the Covid-19 pandemic or indigenous communities in the Global North, skews decision-making and global policy. Similarly, localized examples, like Brussels' daycare data failing to integrate mobility information, highlight the practical limitations of open datasets in addressing accessibility and equity issues.
  5. Legal avenues for redistributing value in the open data ecosystem need to account for the current inequities in open data initiatives. 3/3: Redistributing value in the open data ecosystem must address existing inequities, as open datasets are inherently political and shaped by systemic biases. While principles like FAIR improve data quality and representativeness, they often treat datasets as neutral, ignoring critical gaps. Missing data, such as underrepresentation of Global South countries during the Covid-19 pandemic or indigenous communities in the Global North, skews decision-making and global policy. Similarly, localized examples, like Brussels' daycare data failing to integrate mobility information, highlight the practical limitations of open datasets in addressing accessibility and equity issues.
  6. There should be continued focus on publication of open government datasets by public administration.
  7. Governance mechanisims to support bottom-up initiatives: OD governance mechanisims should also support communities and local environment to interact and benefot of OD ecosystems not just as users od data. Local partnertships and networks might supported by new governance mechanisims with a bottom-up perspective.
2. Circular
  1. Enhanced data creation process (for NGD holders and beyond) through better quality, integration and validation tools.
  2. Enhanced data sharing through better licensing and anonymisation: Important to develop new technical mechanisms and solutions towards the creation of anonymised datasets in an automated way, as well as, the creation of clear legal frameworks and licensing policies that are acceptable to a diverse range of stakeholders, fostering a trustworthy environment for open data exchange.
  3. Technical and contextual support:OD ecosystems should bridge the technical and contextual knowledge gap needed to extract value from data at different stages of its lifecycle.
  4. Dynamic roles and interactions: In circular ecosystems, actors might adopt different roles such as user, providers or intermediaries according to their potential contributions. Interactions are dynamic as well as their role.
  5. Networks and layers: The dinamics and interconnected interactions in OD ecosystems might be represented as networks. A multi layer perspective might contribute to represent different value generation and stakeholders perspectives such as busniness, government and local communities.
3. user-driven
  1. A strong interorganisational culture oriented towards effecting social and economic impact of open data in a participative manner is also useful: "Organisations should have a strong internal culture of contributing to as well as realising value from open data ecosystems. This can then enable inter-organisational cultures, as well as potentially foster partnerships. These organisational cultures can also serve as motivations for actor groups to engage in open data initiatives. Such cultures can be cultivated using formal institutions like legal frameworks, together with informal approaches to creating shared understandings among the various open data actors.
  2. Better interfaces and feedback mechanisms for open data provision: "Important to design and develop the necessary technical mechanisms including new interfaces and user interaction methods that enhance the way NGD holders are creating and sharing open data that is available, usable and valuable to a wide range of stakeholders, fostering a more inclusive and effective open data ecosystem.
  3. Supporting communities of OD Ecosystems: To support the formation of communities around the use of open data, actors of the OD Ecosystem should be knowledgeable of the purposes and practices that can be affected by open data. Shared purposes typically revolve around public and/ or local concerns, therefore they directly affect citizens, local communities or digital communities. Communities of practice typically form when experts, practitioners and academics explore societal problems by developing knowledge, tools, practices that address those problems. Research done in the field of participatory design focuses on empowering communities of shared purpose, while disciplines such as data science and engineering, engineering design, computer science typically form communities of practice around open data.
  4. Ensuring sustainability of OD Ecosystems: In terms of economic sustainability of OD Ecosystems, advocacy for availability of public funds can be accompanied with insights from economic and business models of digital commons/information commons/data commons projects, in particular from collaborative peer production. Contributions to social sustainability can be ensured through the adoption of critically situated approaches to participation from critical data studies.
  5. Considering appropriate legal mechanisms
  6. Partnerships should be formed between public administrations / international organisations and other civic open data user groups to co-develop open data initiatives, which can prompt contributions from other actors.
Key Strategies proposed
1. Inclusive
  1. Open data events should be designed to be accessible by non-specialist users.
  2. The public administration should institutionalize and support open data events
  3. The EU’s Implementing Regulation on High-Value Datasets (2023/138), while fostering innovation by prioritizing economically valuable datasets, needs to be applied more equitably. Critics highlight that focusing on efficiency and market productivity marginalizes citizens' rights and participation. To address this, equitable implementation could involve expanding dataset categories to include social data, as seen in India's focus on poverty alleviation or Glasgow’s project on well-being indicators, ensuring that open data serves not only economic growth but also societal needs and inclusivity.
  4. Legal institutions should focus on ensuring commercial actors in open data initiatives prioritize the public interest. Courts can establish guidelines for open data intermediaries, as illustrated by the ECHR ruling in Satakunnan Markkinapörssi Oy v. Finland, which restricted companies from misusing publicly accessible tax data for sensationalism instead of fostering meaningful public debate. Additionally, public procurement frameworks can be adapted to require private service providers to share data generated during public service delivery in open, machine-readable formats. Barcelona’s use of “data sovereignty clauses” in procurement contracts offers a strong model for balancing commercial involvement with public accountability.
  5. Private law instruments, such as open licenses, can promote equitable value creation in the open data ecosystem by addressing power imbalances and fostering ethical data sharing. While open licenses like Creative Commons and Open Database Licenses were originally designed to counter copyright restrictions and encourage re-use, new challenges like data privacy, competition, and data colonialism necessitate their reimagining. For example, the University of Pretoria’s Data Science Law Lab has developed a data license for African language datasets, requiring stronger share-alike obligations from re-users in developed countries, ensuring value flows back to the Global South and countering data extractivism.
  6. Continued implementation of legal mandates for open government data.
  7. Scoping at local problems and building networks of actors and OD flow among them
2. Circular
  1. Enhance the data creation process by employing several technical mechanisms and technologies to improve critical aspects of data creation, quality and integration, such as augmenting semantic interoperability with the help of Large Language Models to automate or semi-automate processes that improve it, improving metadata annotation in open data portals using AI, or validating/checking technical compliance of data on the Web.
  2. Enhance data's licensing and anonymisation by employing technological facilitators to achieve this purpose, such as Machine Learning, Data encryption, GenAI and legal interoperability facilitators.
  3. Design OD infrastructures as a hubs for creativity, learning, exploration, and knowledge sharing
  4. Focus on value creation and potential contributions showing a rich network of roles and interactions
  5. Visualise OD ecosystems as multi-layer networks mapping
3. User-driven
  1. Fostering Open Data Culture: Cultivating strong organisational cultures oriented towards open data
  2. Enhance the foster improved OD platforms' interface and feedback mechanisms by adopting user-driven design, inclusion of diverse user groups in society and active feedback loops while making use of new technologies wherever this is possible; e.g., Natural Language Processing, Virtual and Augmented Reality, AI, and Chatbots.
  3. Actors should be knowledgeable of the purposes and practices that can be affected through open data.
  4. To ensure sustainability of OD ecosystems, different types of funding models (public funding, private funding, etc.) should be considered.
  5. Use of open licenses should be encouraged
  6. Willingness on part of public administrations to engage in formal and informal partnerships with other civic actor groups.

×

Governance of Open Data

Description:

Major element of open data ecosystem, deals with setting the roles, defining rules of open data initiatives and open data itself, and mechanisms for open data handling (managing open data responsibly, justifiably, and endurably) throughout open data life cycle and beyond. This phase sets the rules in games such as resolving conflicts and power imbalances among stakeholders (data providers, intermediaries, and users), preserving accountability, adopting/developing standards. Good governance leads the open data ecosystem to essential components such as trust, data quality, and FAIR use of open data.

State of the art:

Governance in Open Data Ecosystems (ODEs) is essential for maintaining sustainable data sharing and reuse among stakeholders, including governments, businesses, and individuals (Pollock, 2011; Van Loenen et al., 2021). Traditional governance models have largely followed hierarchical and market-based mechanisms (Chantillon et al., 2017), but recent research highlights the benefits of participatory, commons-based approaches (Gurnstein, 2011; Ostrom, 1990), aiming for greater inclusivity and efficiency. ODE governance must balance regulation, stakeholder involvement, and sustainable data sharing. Hierarchical models rely on strict regulations, ensuring compliance but limiting adaptability (Weerakkody et al., 2017). Market-based governance uses financial incentives to encourage participation but may exclude resource-limited actors (Moore, 1993, 2006). Network-based governance fosters collaboration and trust (Provan & Kenis, 2008) but can suffer from coordination issues (Johnson, 2014a). A growing body of research supports commons-based governance, following Ostrom’s (1990) self-governing principles. This approach promotes boundary-making, shared decision-making, and legal frameworks for sustainability (Davies & Perini, 2016). Commons-based governance facilitates collective management of data resources, improving accessibility and fostering innovation (Cavanillas et al., 2016). Non-governmental data holders face barriers in ODEs, including legal, financial, and technical challenges (Reggi & Dawes, 2022a). Many private entities hesitate to share data due to intellectual property concerns and competitive risks (European Commission, 2020). Uncertainty around data ownership and privacy further complicates participation (Van Loenen et al., 2021). To address these issues, governance strategies must provide regulatory clarity and incentives. Policy interventions such as mandatory data-sharing for publicly funded research or tax benefits for businesses that contribute data can bridge gaps (Weerakkody et al., 2017). Collaborative platforms enable businesses, NGOs, and researchers to negotiate agreements, fostering trust and mutual benefits (Chantillon et al., 2017). Empirical studies indicate that public-private partnerships improve data-sharing practices (Hossain et al., 2016). User engagement is crucial for ODE sustainability, requiring contributions beyond data consumption (Yuan, 2019). Barriers such as limited technical literacy, lack of awareness, and unclear feedback mechanisms hinder participation (Huijboom & Van den Broek, 2011). Motivation theories offer insight into engagement. Self-Determination Theory underscores autonomy, competence, and social connection as key drivers (Ryan & Deci, 2000). Users contribute more when their efforts have meaningful impact and when platforms are user-friendly and rewarding (Wenger, 1998). Fostering communities of practice strengthens engagement by encouraging knowledge-sharing and innovation (Bachtiar et al., 2020). Effective governance should include gamification, recognition systems, and streamlined contribution mechanisms (Davies & Perini, 2016). Open data portals allowing user annotations and dataset updates improve participation (Safarov et al., 2017b). Partnerships with educational institutions and advocacy groups further enhance long-term user engagement (Hossain et al., 2016).

Beyond the state of the art:

Governance frameworks must evolve to address emerging challenges like data sovereignty, AI in decision-making, and the digital divide (Kitchin & Lauriault, 2014). Hybrid models combining hierarchical, market-based, and commons-based approaches can provide regulatory compliance while fostering innovation and collaboration (Micheli et al., 2020; Van Dijck et al., 2018). AI-driven compliance mechanisms can streamline governance by monitoring data-sharing agreements, while blockchain-based smart contracts can enhance transparency (Davies & Perini, 2016; Gurnstein, 2011). However, these technologies must be implemented carefully to prevent exacerbating power asymmetries (Ruijer et al., 2024) and environmental footprint. Ensuring inclusivity in ODE governance remains vital. Marginalized groups often face obstacles in data access and contribution, reinforcing inequalities (Meijer & Potjer, 2018). Policies promoting digital literacy and community-led governance can help address these disparities (Ruijer et al., 2024). Governance strategies must prioritize equity, ensuring open data benefits are widely distributed and decision-making remains inclusive.

Key elements identified
1. Circular
  1. Incentivising open data sharing by open data holders in ODE through governance strategies: An approach to steer the behaviour of non-government data holders towards open data through a governance strategy.
2. user-driven
  1. Coordinating functions should be discharged by public administrations to bring together open data providers and open data users to openly discuss their needs: "Public administrations should routinely consult with open data users and providers from the citizenry, the private sector and the public sector to understand evolving needs of these stakeholders and undertake open data initiatives that respond to these needs. Public administrations should also provide funding to collaborative open data initiatives, which can then serve a coordinating function. At the same time, coordinating functions can also lead to open data ecosystems in unexpected ways. For instance, the UK mapping agency – Ordnance Survey – did create large geographic datasets, but many open data users were dissatisfied that these datasets were not freely distributed. This led to the birth of OpenStreetMap (see https://wiki.openstreetmap.org/wiki/History_of_OpenStreetMap) – an open platform for crowd-source geospatial data, which could be considered an example of a sustainable open data ecosystem. Future research can explore mechanisms for non-governmental actors to discharge coordinating functions.
  2. Encouraging participation and shared decision-making: To encourage polycentricity through participation and collaborative decision making in the OD Ecosystem, actors with more power such as institutions, organizations, communities that represent the status quo should ensure that those typically with less power such as citizens, students, research participants as well as less represented and disadvantaged groups are being actively encouraged to communicate their feedback, needs and concerns, as a first step. Moreover, they should be empowered to actively contribute to the creation of strategies and plans, practices and assessments, products and services. The discipline of (critical) data studies provides tools, approaches and theoretical concepts that challenge existing power structures and propose more just and equitable alternatives.
  3. Partnerships should be formed between public administrations / international organisations and other civic open data user groups to co-develop open data initiatives, which can prompt contributions from other actors.
Key Strategies proposed
1. Circular
  1. Empowering NGOs with Skills and Resources: 1. Promote training to improve technical skills for NGOs that lack them. 2. Creating NGOs' common platform for open data sharing. 3. Financial resources are often scarce, so to support NGOs' projects long-term, more financial cooperation is needed. 4. Promote training to improve open data awareness for NGOs that lack it.
2. user-driven
  1. Coordinating functions to be discharged by governmental actors. With the right incentives, such functions can be discharged by non-governmental actors as well"
  2. Powerfull actors and organisations should ensure that less powered , less represented and disadvantaged actors and organizations are encourage to provide feedback and even actively contribute.
  3. Willingness on part of public administrations to engage in formal and informal partnerships with other civic actor groups

×

Value of Open Data

Description:

Well-organized open data initiatives generate values of different types, such as tangible and intangible. These values can be growth in the economy, social awareness and empowerment, and sustainability in the field from which this open data is generated and shared. The value generation and its fair distribution is a long-standing discussion topic in open data. Open data ecosystems also talk about fair value distribution among the involved stakeholders as per their contributions in open data initiatives. As the value creation is dependent on several other highly dynamic factors such as stakeholders’ interest, skills, regulations, data quality, and OD infrastructure.

State of the art:

The current state of the art in value generation within Open Data Ecosystems showcases the balance between the autonomous, self-interested actors and the collaborative mechanisms that enable value creation. Davies (2011) emphasizes that successful value generation requires mobilizing a combination of technical, social, and political resources while fostering coordination and self-organization among ecosystem participants (Davies, 2011). Csáki (2019) further elaborates on ODEs as dynamic systems where actors and groups create shared meaning and value around open data, with their interactions shaping the ecosystem’s growth and health (Csáki, 2019). Oliveira and Lóscio (2018) identify four core elements of an ODE—resources, roles, actors, and relationships—underscoring actor’s flexibility in adopting various roles (Oliveira & Lóscio, 2018). Strategies for maximizing value are shifting from top-down interventions to community-driven approaches, recognizing the inadequacies of hierarchical models for handling multidirectional data flows.

Beyond the state of the art:

While current research focuses on value creation in the open data ecosystem, there is a need to deepen the knowledge and understanding of value distribution mechanisms and their adaptability in the dynamic environment of an open data ecosystem. The development of value distribution algorithms in response to emerging stakeholder scenarios is crucial. Additionally, the creation of feedback loops and ecosystem health metrics can help facilitate value creation and fair distribution within the open data ecosystem. Furthermore, introducing new circular business models into the open data ecosystem and exploring how they can integrate with the circular economy and multi-stakeholder collaborations will contribute to more resilient and adaptable open data ecosystems.

×

AI for Data

Description:

Artificial intelligence is considered as a driving force to drive the synergies in the field of open data. Its purpose is two-fold, firstly, using AI to improve the open data at various stages such as quality, metadata, alignment with standards, etc. on the other hand use of AI to develop applications and use-cases using the open data to bring innovation and value-generation. AI4Data means improving open data in terms of quality (not limited to only quality, interoperability, integration, metadata generation, recommendation etc.). Data4AI means using open data as a source and bringing AI analytical capabilities to foster innovation. Data4AI covers the aspects of making open data available for AI models training and testing. Advancement in AI4Data and Data4AI can open new opportunities in data-driven solutions or easy data product developments.

State of the art:

Artificial Intelligence has revolutionized how open data is used, whether in data quality enhancements or innovative applications. The AI4Data paradigm focuses on enhancing open data at various stages, such as generating or improving metadata, enhancing data quality, making data interoperable, and standardizing it. Several studies have explored these topics, and they are still being investigated avidly following the advent of more advanced models. For instance, Ahmed et al., (2024) created a hybrid method (BRYT) combining conventional and advanced methodologies such as BERT, RAKE, YAKE, TextRank to extract keywords for open data sets. Song et al., (2024) utilize contemporary LLMs such as LLAMA and Mistral to extract metadata from more descriptive fields inside the dataset. Turgunbaev (2024) discusses various conventional machine learning methodologies for the particular usecase of metadata extraction. Giglou et al., (2024) explore state of the art LLMs such as GPT, Llama, Mistral to perform ontology mapping ultimately as a use case for data interoperability. Bakker & Scala, (2024) perform ontology learning from textual data using contemporary LLM: GPT-4o. Alharbi et al., (n.d.) employ LLM- GPT to create an ontological structure – Knowledge Graph from medical texts. Maratsi et al., (2024) explore multiple ontologies and their possible interconnected mapping. Moreover, they employ GPT 3.5 and GPT 4 to generate the mappings and validate its accuracy on manual data. Data4AI paradigm focuses on utilizing open data to enhance the development of AI applications. Open datasets provide significant opportunities for training and improving machine learning models. Access to high-quality open data speeds up AI research by lowering the costs and time required for data collection and preparation. Although it has a lot of room for research, few studies have explored the idea. In one of such studies, Weber et al., (2024) emphasize the significance of open datasets for training machine learning models. They reflect the challenges associated to transparency, data access, and data curation, crucial for advancing open-source AI development. Arbeláez et al., (2024) underscore the significance of employing open data which might be diverse and even unlabeled as a better cost-effective alternative to custom-labelled data which might be expensive and might hinder development. Alam et al., (2023) highlight the role of Open Data and Explainable AI in enhancing transparency, accountability, and interpretability in healthcare AI models. Employing diverse open data, the models can achieve better reliability and accuracy, promoting trust and openness. State-of-the-art reflects how the synergy between AI4Data and Data4AI can create robust feedback loops. Improved data quality and metadata quality (AI4Data) lead to greater model performance, while AI applications and innovations (Data4AI) can further enhance data collection methods through the use of tools using AI. This dual advancement opens opportunities for new data-driven solutions, ranging from automated decision-making systems to scalable data products that address societal challenges in areas like healthcare, climate change, and smart cities.

Beyond the state of the art:

The future holds significant potential for AI4Data and Data4AI. It would revolutionize data curation and AI model training. The current state particularly focuses on improving data quality, interoperability, and accessibility. Moving forward, we can also focus on self-optimizing dynamic ecosystems that use AI and reinforcement learning to maintain the integrity of data, metadata, and interoperability standards. From conventional static systems, it would move to more dynamic systems that would be self-sufficient in addressing changing contexts and user needs. For instance, whatever data-related needs are communicated to the systems via LLMs, which understand the context of user needs, it would trigger the pertinent sub-model to address it and resolve it. Moreover, data in ontological structure and knowledge graphs could prove to be a game changer for transparency, interoperability, and, ultimately, for training models. We could leverage the semantic capabilities of Large Language Models to build domain-specific ontologies to structure open data. It would highly enhance the interoperability and openness of the models that will be trained through those. The domain-specific ontologies and structure would highly enhance the AI model training for any particular domain task. Apart from being accurate it would turn out to be highly transparent and explainable in terms of its predictions. In the future, we may also see the emergence of AI-driven data marketplaces that facilitate the dynamic exchange and monetization of open data through blockchain-based platforms. These platforms would potentially enable data sharing by ensuring provenance, secure access, and fair compensation. This approach would significantly enhance data-driven innovation and promote economic growth on a global scale.

Key elements identified
1. Circular
  1. Enhanced data creation process (for NGD holders and beyond) through better quality, integration and validation tools.
Key Strategies proposed
1. Circular
  1. Enhance the data creation process by employing several technical mechanisms and technologies to improve critical aspects of data creation, quality and integration, such as augmenting semantic interoperability with the help of Large Language Models to automate or semi-automate processes that improve it, improving metadata annotation in open data portals using AI, or validating/checking technical compliance of data on the Web.
Technical Solutions
1. BRYT: Automated keyword extraction for open datasets (Link: https://www.sciencedirect.com/science/article/pii/S2667305324000954)
2. Towards Cross-Domain Linking of Data: A Semantic Mapping of Cultural Heritage Ontologies (Link: https://dl.acm.org/doi/pdf/10.1145/3657054.3657077)

×

Feedback Evaluation and Redefinition

• Description:

Collecting feedback and evaluating the open data processes in open data ecosystems are necessary to redefine the processes, requiring modifications to cope with the challenges related to open data ecosystems. The feedback mechanisms are helpful in identifying the gaps, refining strategies/policies, and adapting to the varying external/internal needs. Additionally, feedback mechanisms are crucial for evaluating the effectiveness of the open data portals in promoting the quality of data and user participation. To complete the feedback loop, evaluation frameworks are critical for quantifying the results and impacts and aligning efforts with the end objectives. For long-term success and sustainability, robust and iterative feedback mechanisms, evaluation and redefinition of processes are vigorous.

State of the art:

Herrera et al. (2023) argue that achieving user-centered evaluation hinges on the strategic adoption of the mental model concept (Herrera-Murillo et al., 2023). Mental models enable individuals to construct descriptions of a system’s purpose and structure, explain its functionality and observed states, and anticipate future behaviors. Designers, who often possess intricate mental models of their own creations, may assume that every feature is intuitive and easy to grasp. However, users typically operate with more limited mental models, making them more prone to errors and perceiving the system as significantly more complex. The disparity between designers’ and users’ mental models frequently leads to usability challenges. By systematically identifying and bridging these gaps through evaluation, a truly user-centered open data ecosystem can be realized. The open data portals have integrated web forms or specialized discussion forums in order to encourage communication among the various user communities. These forums allow users to share their experiences regarding how they have utilized the data that is made available through the open data portals. Due to the significant degree of variability that exists in this kind of user feedback, it is not possible to automatically compare the input that comes from different attempts (Aziz et al., 2022).

Beyond the state of the art:

The increasing incorporation of artificial intelligence tools to enhance the usability of open data can play a crucial role in bridging misalignments between users’ mental models and the complexity of datasets. AI-driven personalization can dynamically adapt open data portals to different expertise levels, offering tailored recommendations and adaptive interfaces. Additionally, cognitive feedback mechanisms, such as real-time clarifications and interactive tutorials, can help users navigate and interpret data more effectively. Conversational AI agents can further support understanding by providing dataset explanations, suggesting analytical approaches, and guiding users in their exploration. Moreover, AI-powered error detection can improve data quality by identifying inconsistencies, biases, or anomalies, increasing trust in open data sources. AI-generated data summarization techniques, whether textual or visual, can also facilitate mental model alignment, making complex datasets more accessible and actionable. By integrating these AI-driven solutions, open data ecosystems can evolve into more user-centered environments, democratizing access and empowering a wider range of users to engage with data effectively.

Key elements identified
1. Circular
  1. Promoting reusable dataset integration techniques: Techniques like data blending and augmentation improve the ecosystem’s efficiency and user engagement.
2. user-driven
  1. User-driven Feedback mechanisms: Feedback mechanisms, including dataser rating systems, dataset comments, and direct user feedback through discussion forum, are essential for enhancing data quality and ensuring user-friendly portals.
Key Strategies proposed
1. Circular
  1. 1. Provide training on dataset linking and blending tools 2. Enable version control for community-shared datasets.
2. user-driven
  1. Integrates user contributions so that it can improve overall dataset relevance and usability. Establish effective feedback mechanisms.

×

The Open Data Infrastructure Layer

Description:

The technological components of open data ecosystems , including tools, standards, services, and software applications, support data storage, sharing, and operations such as data integration and interoperability across diverse data domains, forming the infrastructure of open data ecosystems. OD infrastructure reinforces the entire open data ecosystem by enabling smooth data sharing and supporting collaborative activities. OD infrastructure faces challenges with the advancement of technologies, such as existing usage of obsolete technologies and unstandardized practices, which hinders its effectiveness. It's worthy in the OD ecosystem to struggle with modern, scalable, and interoperable open data infrastructure for a sustainable OD ecosystem.

State of the art:

Several studies emphasize the importance of scalable and interoperable infrastructures to ensure the sustainability of OD ecosystems. Kirstein et al., (2020) describe the idea of large-scale open data management platforms based on semantic web technologies. Conde et al., (2024) further explore metadata-driven integration of open data portals into data spaces, advocating for automated metadata generation and validation through standardized models. Grossman et al., (2016 propose the data commons model, which co-locates data, storage, and computational resources with common analysis tools, emphasizing persistent digital identifiers, metadata services, APIs, and data portability for interoperability. Collectively, these studies underscore the need for modern, scalable, and standardized OD infrastructures that leverage semantic technologies, metadata-driven integration, and interoperable frameworks to enhance the effectiveness and long-term viability of OD ecosystems.

Beyond the state of the art:

The future of open data ecosystems could embrace adaptive and self-optimizing data infrastructures that evolve in response to real-time conditions and user interactions, what we could define as living open data infrastructures. Unlike traditional open data repositories, which rely on periodic updates and manual interventions, living open data infrastructures would incorporate machine learning, automation, and feedback loops to enhance data quality, relevance, and interoperability dynamically. By integrating AI-driven mechanisms, open datasets could learn from usage patterns, detect inconsistencies, and refine their structure and content based on community contributions and contextual changes. This approach would foster greater responsiveness, accuracy, and efficiency, ensuring that open data remains continuously updated and aligned with user needs. Moreover, self-optimizing datasets could enhance cross-domain interoperability by adapting to evolving data standards and seamlessly integrating new information sources. However, realizing this vision requires addressing key challenges, such as ethical considerations, transparency in algorithmic decision-making, and safeguards against biases in automated learning processes. Additionally, governance frameworks would need to balance openness with reliability, ensuring that self-optimizing datasets remain verifiable and accountable to the broader OD community. By embracing living open data infrastructures, future OD ecosystems could move beyond static data repositories, enabling a more dynamic, intelligent, and sustainable approach to data sharing and collaboration.

Key elements identified
1. Inclusive
  1. Shared open infrastructures for publishing and use of data and information are important, including for non-governmental data holders
  2. Community-centric design:Students' motivations highlight the importance of focusing on real-world relevance and fostering their contributions.
  3. The availability of open data platform for non-government open data provision and reuse.
2. Circular
  1. Open Data accessibility and interoperability: The Open Data Platform (ODP) requires backing for data integration and interoperability from a range of stakeholders, including governmental and non-governmental entities.
  2. Scalability and performance of the portal: the open data volume, variety and velocity increases by the time, so open data portals should be scalable, and performance should be high to accommodate these variations over the time.
  3. Enhanced data creation process (for NGD holders and beyond) through better quality, integration and validation tools.
  4. Boundary objects:The concept of shared tools or platforms as "boundary objects" facilitates communication and collaboration across diverse user communities.
  5. User-friendly interfaces Establish an easy-to-navigate interface for the datasets and analysis tools that anyone can understand and use.
  6. Enhancing interoperability through open standards and metadata: Ensuring technical openness in datasets through interoperable standards improves system integration and enhances usability
3. user-driven
  1. Better interfaces and feedback mechanisms for open data provision: "Important to design and develop the necessary technical mechanisms including new interfaces and user interaction methods that enhance the way NGD holders are creating and sharing open data that is available, usable and valuable to a wide range of stakeholders, fostering a more inclusive and effective open data ecosystem.
  2. Identification of gaps between the needs of user groups and the current features of open data platforms with respect to user interfaces
  3. FAIR Principles: open data portal must be upto par of FAIR data principles for all users. It suppots, FAIR principles for specialist and non-specialist open data stakeholders
  4. Smooth intergration of open data from diverse domains: Open data publishers requires a smooth data integration process to combine data from diverse sources to make a collective decision. In this way, several issues with the open data portals such as data interoperability, data transformation, data harmonization and intergration of tools and technologies can be achieved.
  5. Multi-format data access: provide the end-users with open data in multi-formats such as open formats, APIs, bulk downloads to tackle diverse needs
  6. Unified view for open data publication process Open Data providers requires unified view of publishing dashboad. For instance, several government levels participates in open data publications, regional, decentralized, and muncipal level. They should have access to unified portal to publish their datasets
  7. "Anonymisation of the open dataset before publishing must be the norm" Open Data providers must have access to tools to increase the anonymity of open data before making it public.
  8. Making available integrated datasets from diverse fields for users with diverse needs is needed: The necessity for integration of data and accessibility feasible for user groups with varying needs (e.g., visual impairment, dyslexia, and others)
  9. Portability of infrastructure in the open data domain: The portability of infrastructure will help transfer the knowledge of OD infrastructure to another organisation
  10. Community Portal for interaction "The data portal should provide a place where users can connect with other users in a community,based on similar interests.
  11. Designing an ecology of interoperable projects To focus on interoperability and data portability and have a broad understanding of these concepts. In particular, efforts for interoperability should encompass technical interoperability (through, for example, semantic and syntactic interoperability of open data systems/portals and through standardised formats for (open) data, as also noted in the empirical data collected) as well as generative interoperability (through adoption of policies aimed at nurturing public spaces for decision-making in relation to open data needs and challenges). Support should be provided to regulatory measures aimed at broad interoperability and portability, through advocacy and political action.
4. Skill-based
  1. Developing low-code tools for educators and data intermediaries
Key Strategies proposed
1. Inclusive
  1. Public financing and maintenance of shared open infrastructures, like the European Open Science Cloud
  2. Community-centric design prioritizes the needs, preferences and involvement of communities in the open data ecosystem.
  3. Establish a national or supra-national consortium for open data sharing by non-public organizations. Among other things, this consortium could develop and maintain non-proprietary technical infrastructure for open data sharing. The consortium may be governed by a committee composed of public, private, and civil sector stakeholders that facilitate collective decision-making on various technical and non-technical aspects. The consortium and its infrastructure may be supported by joint multi-stakeholder financial cooperation based on agreed terms.
2. Circular
  1. Employing Open Data Standards and Formats for Ecosystem Circularity: Employing open data standards and formats can streamline data accessibility and sharing within open data ecosystems, enhancing their circularity.
  2. Leveraging Big Data and Cloud Computing for Scalability and Performance: Follow the big data and cloud computing technologies to make the portal more scalable and improve performance.
  3. Enhance the data creation process by employing several technical mechanisms and technologies to improve critical aspects of data creation, quality and integration, such as augmenting semantic interoperability with the help of Large Language Models to automate or semi-automate processes that improve it, improving metadata annotation in open data portals using AI, or validating/checking technical compliance of data on the Web.
  4. Boundary objects are essential for fostering inclusivity, enhancing collaboration, and bridging gaps between stakeholders such as governments, NGOs, businesses, and individual citizens.
  5. Facilitate better data exploration and comprehension by including interactive visualisation tools.
  6. 1. Adopt open standards for datasets 2. Implement metadata validation tools for consistent data quality
3. user-driven
  1. Enhance the foster improved OD platforms' interface and feedback mechanisms by adopting user-driven design, inclusion of diverse user groups in society and active feedback loops while making use of new technologies wherever this is possible; e.g., Natural Language Processing, Virtual and Augmented Reality, AI, and Chatbots.
  2. The data should be findable to wide range of users with diverse searching options.
  3. The FAIR Principles should be applied to achive the user-driven open data ecosystem.
  4. Open Data ecosystem should facilitiate the intergration of open data from cross-domains (either government, non-government, NGOs).
  5. Develop a standardized open data portal interface to make the data publication process easier.
  6. Use/develop tools to anonymize the open data
  7. Introduce tools for unified view of data and physically impaired user groups.
  8. Seek guideness and infrastructure from the already developed portals from other countries, areas, regions to develop your one.
  9. Open Data community portal development helps in identification of issues with data, the possible use of data, future improvements.
  10. Focus should be put on - different types of - interoperability and data portability
4. Skill-based
  1. "1. Create and distribute low-code tools for data analysis. 2. Offer training to educators and intermediaries on using these platforms.

×

The Open Data Ecosystem

Description:

The open data ecosystem is considered pivotal for sustainable value creation through open data utilization. In this study, it is used as an umbrella term, as shown in Figure 5, and the most outer bounding rectangle covering all major aspects, which plays a role directly or indirectly in the open data ecosystem's sustainable value creation. The OD ecosystem is comprised of a dynamic environment covering aspects related to open data stakeholders, driving forces, value creation, governance, and OD infrastructure. The motive behind creation of the open data ecosystem is to drive all these components smoothly and efficiently for the innovation and value creation across different OD domains. In OD ecosystems, stakeholder groups work collaboratively to stimulate the potential of OD. Despite the potential of open data (OD) ecosystems, they face several challenges regarding the standardization of tools, techniques, and methods for implementing OD initiatives, as well as generating value through the use and reuse of OD based on needs and requirements. Addressing these challenges can enhance value creation and support the sustainability of the OD ecosystem.

State of the art:

Open Data ecosystem is a concept for public, private, and non-profit organizations, several studies have explored its relevance for innovation, sustainability, value-creation and transparency (Styrin et al., 2017; Van Loenen et al., 2021; Zuiderwijk et al., 2014) . Open data ecosystems can be compared to open-source software, but in ODE, data is shared under public licenses instead of software (Runeson et al., 2021). open data ecosystems are more about data sharing under public licenses, whereas open-source software refers specifically to sharing software code. It should be noted that Open-Source Software depends on the data, particularly since surge in artificial intelligence and generative intelligence models has been witnessed. Open data ecosystems involve several types of actors (commercial, non-profit, researchers), as study by Runeson et al. (2021) have explored ODE for commercial actors. ODE concepts are at very beginning stage, where they are facing challenges with the proper actor inclusion and roles, value-creation, value-distribution, user-needs elicitation, quality of data, utilization of artificial intelligence, adding value to the Open data, and making all these steps transparent within open data ecosystems and beyond. Open data systems face challenges to achieve value-creation and sustainability due to supply and demand mismatch, exclusion of certain user groups and domains, linear value-creation, and lack of skills have been explored at domain and country level such as finance, legal, education, geographic, and agricultural (Van Loenen et al., 2021). Open data impacts and its realization depends on stimulation of technical, social, and political resources, beyond the data supply but also making a contribution chain of activities around the datasets (Davies, 2011). The ODE complexity and challenge space observed from its multifaceted and multiprofessional nature and even gets multiplied when heterogeneity from the data, domains, and actors and each of them having their own specific challenges at different spatial and temporal levels gets into this.

Beyond the state of the art:

In this deliverable , we have examined the diversified nature of open data ecosystem concepts and associated challenges with them, which suggests a clear need for the open data ecosystem framework creation — containing all the major components. The role of ODE framework’s components in stimulating the ODE drivers. Furthermore, proposing and formalizing strategies for each ODE framework’s components, so that actual ODE can be sustained and value-creating for all involved actors. Also new waves of AI such as open-source availability of AI models (e.g., large language models) can give and take benefits from the open data ecosystems. Additionally, emerging AI technologies, such as the open-source availability of AI models (e.g., large language models), can both benefit from and contribute to open data ecosystems. The formalization of ODE frameworks and corresponding strategies would be one step towards achieving sustainability and value-creating open data ecosystems.

Key elements identified
1. Inclusive
  1. Lowering barriers to data access Ensuring the inclusivity of open data portals involves the provision of multilingual support, user-friendly interfaces, and accessibility for non-technical users.
  2. Addressing barriers to local community access to open data: Communities require localized, interpretable datasets accompanied by metadata to utilize data effectively for decision-making and advancements in society.
Key Strategies proposed
1. Inclusive
  1. Simplify interfaces with user-friendly designs and incorporate multilingual support and accessibility features.
  2. 1. Provide tools for local data interpretation 2. Create community-specific manuals and localized data insights platforms

Data Provision

Driving Force: Data Quality

Description:

State of the art:

Beyond the state of the art:

Users stakeholders groups

Data for Artificial Intelligence

Description:

State of the art:

Beyond the state of the art:

Producer's stakeholders Groups

Description:

State of the art:

Beyond the state of the art:

Intermediaries

Description:

State of the art:

Beyond the state of the art:

User Stakeholders

Description:

State of the art:

Beyond the state of the art:

Data Provision to Data Usage

Data Exchange

Description:

State of the art:

Beyond the state of the art:

Value Co-Creation/Impact/Innovation

Description:

State of the art:

Beyond the state of the art:

Value Co-Creation, Impact, and Innovation

Collaboration and Network Effect

Description:

State of the art:

Beyond the state of the art:

Value Arrows

Incentive Arrows

Description:

State of the art:

Beyond the state of the art:

Usage to Provision

Open Data Ecosystem Drivers

Overll Challenges

ODECO Research Gaps

Overll Strategies

User-Driven Strategies

Circular Strategies

Inclusive Strategies

Recommendation

Technical Recommendations

Governance Recommendation

Resources

Description:

State of the art :

Beyond the state of the art:

Enabling Environment

Description:

State of the art:

Beyond the state of the art:

Governance of Open Data

Description:

State of the art:

Beyond the state of the art:

Value of Open Data

Description:

State of the art:

Beyond the state of the art:

AI for Data

Description:

State of the art:

Beyond the state of the art:

Data Exchange 2

Feedback Evaluation and Redefinition

• Description:

State of the art:

Beyond the state of the art:

The Open Data Infrastructure Layer

Description:

State of the art: