PowerPoint Presentation

[Audio] Welcome to module 3 "Data, platforms and interoperability".

[Audio] In Module 2 we followed the product along the chain and saw digital technologies at every stage. In this module we focus on what they all produce and exchange: data. Data is the thread that connects the actors, and the rules for how it flows will largely decide who benefits from digitalisation. Let us start with where data comes from. On the farm, there is data from fields, machines and animals: soil measurements, yield maps, application records, milk yields, feed intake. There is environmental data: weather stations, forecasts, satellite imagery. There is market data on prices and demand, logistics data on shipments and conditions, and consumer data on purchases and preferences. Some data is held privately by farms and companies, and some is public or open, such as the Copernicus satellite data we discussed. A useful way to think about it is the data value chain: data is collected, stored, integrated with other data, analysed, used for a decision, and sometimes shared with others. Each step adds value, and each step involves different actors. A farmer may collect soil data, a technology company may store and analyse it, an advisor may turn it into a recommendation, and a processor may later want to use the same information to plan its operations. The most important point is that value often emerges from combination. Soil data alone is useful. Combined with weather forecasts, machinery records and market prices, it can support much better decisions. But combination requires that data from different owners can be brought together, and that is exactly where questions of ownership, trust and compatibility arise. It also helps to distinguish types of data. Personal data, meaning data relating to an identifiable person, is protected under the General Data Protection Regulation. Non-personal data, such as soil moisture readings, is not covered by that regulation, but it may still be commercially sensitive. And the line can be thinner than we think: a review of the literature on data ownership in agriculture points out that something like a tractor's GPS track or field boundaries can count as personal data if it directly or indirectly identifies a person. Remember what Wolfert and colleagues found in their review: big data in farming is expected to shift roles and power relations in food chains. So we can organise the rest of this module around four simple questions. Who generates the data? Who holds it? Who can use it, and under what conditions? And who benefits? Pause for a moment. For your own product example, list three data sets that exist somewhere in the chain, and write next to each one who holds it today..

[Audio] "Who owns the data?" is probably the most frequently asked question in digital agriculture, and the honest answer is more complicated than people expect. For personal data, European law gives individuals clear rights. For non-personal data, such as machine or sensor data, there is no general property right in data. What exists in practice is a combination of contracts, technical control and, increasingly, specific access rights. That is why experts often prefer to talk about control, access and use rather than ownership. The practical question is not "who owns it" but "who can access it, who can use it, for what purposes, and who decides?" Research on farmers' attitudes shows why this matters. Studies on farmers' reluctance to share data point to concerns about who will see the data and how it may be used. Farmers often feel that data about their inputs, decisions, yields and accounts belongs to their own business. They worry that technology companies could reuse it to build other services, or to influence prices, while the farmer is excluded from the benefits. A key paper by Wiseman, Sanderson, Zhang and Jakku examined this through the lens of the laws affecting smart farming. Other work notes that farmers can feel vulnerable when negotiating contracts with large technology providers, whose licensing terms may sit under foreign jurisdictions. And since much of digital farming is done through online terms and conditions, many farmers never read, or fully understand, what they agree to. This is an important point about power. Earlier we talked about asymmetries between small farms and large retailers. The same asymmetry appears here between individual farmers and large technology and machinery companies. The Food and Agriculture Organization's 2022 report on agricultural automation warns that if technology companies keep and own the data, this could favour already powerful actors and may lead to data monopolies. What builds trust? Research by Jakku and colleagues on trust in smart farming highlights transparency, meaning clear information about what data is collected, why, and with whom it is shared, and benefit-sharing, meaning that those who generate the data also receive something in return, whether better services, lower prices or a share of the value. For training and advisory work, there is a practical lesson. Farmers and cooperatives need to be able to ask the right questions before signing up for a digital service: what data is collected, who can see it, can I take it with me if I leave, can it be used for purposes other than my service, and what happens if the company is sold? Think about your own context. If a cooperative asked you for a one-page checklist of questions to ask a technology provider about data, what would be the first three items?.

[Audio] How can we govern data sharing in practice? Europe has taken two complementary routes: a voluntary route led by the sector, and a legislative route. The voluntary route is the EU Code of Conduct on agricultural data sharing by contractual agreement. It was launched in Brussels on 23 April 2018 by a coalition of associations from across the agri-food chain, including the farmers' and cooperatives' organisation Copa-Cogeca and the machinery industry association CEMA, along with associations from the fertiliser, crop protection, seed, feed and livestock breeding sectors. The Code is non-binding. It sets out general principles and a contract checklist, and it keeps the farmer at the centre of the collection, processing and management of the data. Its key principles cover the attribution of rights over the data to the data originator, access, control and portability, and data protection and transparency. It focuses mainly on non-personal data. The Code has been valuable in establishing a common language. But as a voluntary document it has limits. Commentators note that conditions are written into individual contracts, which in practice can mean many complex and unclear contracts, and nothing obliges a company to follow the Code. The legislative route is the Data Act, Regulation 2023/2854. Most of its provisions became applicable on 12 September 2025. It establishes a user-centred regime of access and sharing for data generated by connected products and related services, which includes industrial machinery, so many modern connected agricultural machines may fall in scope, although the application to particular cases should be checked. Users have a right to access the data their products generate. Companies that hold the data may not use it, in the case of non-personal data, without a contract with the user. From 12 September 2026, newly placed products must be designed so that data is accessible to the user by default; that date has just passed. And the Act includes rules on unfair contractual terms in business-to-business data contracts, which apply to contracts under the original timetable from 2025, and to certain older long-running contracts from 2027. For farmers, this is potentially significant, because it moves some of the principles of the Code of Conduct into binding law. However, rights on paper are only the start. Using them requires awareness, technical means to retrieve and reuse the data, and the capacity to negotiate. Cooperatives, advisors and trusted intermediaries have an important role in helping farmers exercise these rights. Note also that this area is moving quickly, so anyone using these slides in teaching should check the latest status of the legislation..

[Audio] Let us now look at a set of principles that you will meet in almost every European research project: the FAIR principles. FAIR stands for Findable, Accessible, Interoperable and Reusable. They were published in 2016 in the journal Scientific Data by Mark Wilkinson and a large group of colleagues from academia, industry, funding agencies and publishers. The idea was to give a concise, measurable set of guidelines for improving the reuse of data. A distinctive feature of the principles is that they stress the ability of machines, not only humans, to find and use data automatically. In practical terms, "Findable" means that data and its description, the metadata, carry persistent identifiers and are registered somewhere searchable, so that someone can discover them. "Accessible" means that once found, it is clear how to access the data, which may include authentication or authorisation. "Interoperable" means that the data uses shared formats and vocabularies so that it can be combined with other data. "Reusable" means that the data comes with a clear usage licence and information about its provenance, meaning where it came from and how it was produced, so that others can judge whether and how to use it. Two clarifications are important. First, the authors emphasise that the principles are guidelines that come before implementation. They do not prescribe any specific technology or standard, and they are not themselves a standard. So "FAIR" is not a product that you can buy, but a direction of travel that can be implemented in many ways. Second, FAIR does not mean open. Data can be FAIR and still be restricted: for example, a company's private farm data can be well described, accessible only to authorised partners, and licensed for specific uses. This is especially important in agriculture, where much of the valuable data is commercially sensitive or personal. A common phrase is that data should be "as open as possible, as closed as necessary". For our topic, FAIR matters in three ways. In research and education, it allows datasets from experiments, pilots and living labs to be reused by others, which is a strong argument for teaching data management to students. In advisory services, well-documented data allows tools from different providers to work together. And in shared platforms and data spaces, which we will discuss shortly, FAIR-like practices are the foundation for making data from different sources work together. A good example of the kind of resource that supports FAIR in our sector is the FAO's AGROVOC vocabulary, which we will see in the next slide. For reflection: think of a dataset you or your organisation holds. How would you rate it on each of the four letters, and what is the cheapest improvement you could make?.

[Audio] Interoperability means the ability of different systems and organisations to exchange data and to use it meaningfully. In practice it is one of the biggest obstacles in digital agriculture: a farmer may use machinery from one manufacturer, software from another, a weather service from a third and a cooperative platform from a fourth, and getting data to flow between them can be painful. A helpful framework comes from the European Interoperability Framework, published by the European Commission in 2017. Although it was designed to guide public administrations in delivering digital public services, its core idea applies well to our sector: interoperability has four layers, and it is not only a technical matter. The first is legal interoperability: organisations operating under different legal frameworks must be able to work together. The second is organisational: agreed processes, roles and responsibilities. The third is semantic: a shared understanding of the meaning of the data. The fourth is technical: the interfaces and protocols that move the data. The framework also stresses governance across all the layers. Agriculture offers excellent examples at the technical and semantic levels. At the technical level, ISOBUS, based on the standard ISO 11783, has become the de facto standard for communication between tractors and implements from different manufacturers. At the semantic level, AGROVOC, a multilingual vocabulary maintained by FAO, standardises the terms for agriculture, food and related fields and is published as linked open data, which helps data from different sources and languages to be classified in a consistent way. But the ISOBUS story also teaches an important lesson. The Agricultural Industry Electronics Foundation, AEF, explains that even though manufacturers agreed on a global standard, not all machines turned out to be compatible, because the ISO standard was not specific enough in several places. The AEF therefore introduced additional guidelines and a conformance test, so that products that pass can carry a label showing that they work together. In other words, a standard is necessary but not sufficient. You also need detailed specifications, testing, certification and a governance body that the industry trusts. Why does this matter for farmers and for the value chain? Without interoperability, farmers face lock-in, because switching providers means losing data or functionality. Companies duplicate effort building one-off connectors. And data stays in silos, which prevents exactly the combination of sources that creates value. The solution is rarely a single global standard. More often it is a combination of standards for the most important interfaces, shared vocabularies, translation layers between systems, and agreements among actors. We will see in a few slides how European data spaces try to put this together..

[Audio] We now come to platforms, which are central to the ecosystem concept we introduced in Module 1. A platform is a system that connects different groups of participants and allows them to interact, and often allows third parties to build services on top of it. Economists describe platforms as operating in two-sided or multi-sided markets. A classic reference is the 2003 paper by Jean-Charles Rochet and Jean Tirole in the Journal of the European Economic Association, which models how platforms that connect two groups of users compete. The key insight is the presence of network effects: the value of the platform to one group increases when more participants from the other group join. A farm management platform becomes more attractive to farmers when more service providers offer apps on it, and more attractive to service providers when more farmers use it. In agriculture, researchers from Wageningen have applied the idea of software ecosystems to farming. In a 2016 paper in the journal Computers and Electronics in Agriculture, Kruize and colleagues proposed a reference architecture for farm software ecosystems, intended to map, assess, design and implement them. Follow-up work describes the typical actors in such an ecosystem: the software vendor, the agricultural service provider, the agri-food company, and the infrastructure provider. This is a very practical way to analyse a real situation: who plays which role, who connects to whom, and who controls the interfaces. What opportunities do platforms bring? They can lower entry barriers: a small start-up can offer an irrigation advisory app without building a complete farm management system, because it can plug into an existing platform. They can integrate data from several sources in one place for the farmer. And they can create a marketplace for services. What risks should we watch? First, lock-in: the more a farmer's data and workflows live on a single platform, the harder it is to leave. Second, concentration of power: the platform owner sets the rules, fees and data terms, and can see what happens across the whole ecosystem. Third, dependency: the farmer's operations may rely on a company whose strategy can change. Therefore, the governance of platforms matters as much as their technology. Important questions include whether data can be exported in a usable format, whether third-party developers get fair access, how fees are set, and whether the platform is operated by a commercial company, a cooperative or a public body. Different ownership models lead to very different incentives. For discussion: would you prefer your cooperative to build its own platform, join a commercial one, or use a neutral public infrastructure? What would you want in the rules?.

[Audio] Let us put the pieces together by looking at the European policy approach: data spaces. In February 2020, the European Commission adopted the European Strategy for Data. Its aim is a single market for data, in which data can flow across sectors and countries under common rules and European values. A central tool is the development of common European data spaces in strategic sectors. The original strategy listed nine, including industry, the Green Deal, mobility, health, finance, energy, agriculture, public administration and skills. According to the Commission, common data spaces are now being developed across fourteen sectors or domains. What is a data space? It is not a single database or platform. It is a framework in which organisations can share data under common governance rules, with agreed technical specifications and trust mechanisms, while each participant keeps control of its own data. The approach is intended to be decentralised. Several legal instruments support this. We already discussed the Data Act. The Data Governance Act, Regulation 2022/868, has applied since September 2023. Among other things, it regulates data intermediation services, which are neutral third parties that connect those who hold data with those who want to use it, with requirements designed to make them trustworthy organisers of data sharing. It also creates a framework for data altruism, meaning the voluntary sharing of data for the public interest. For agriculture, data cooperatives and neutral intermediaries are an interesting model. Please note that the Commission has proposed changes to this legal landscape, including a 2025 proposal to fold the data governance rules into the Data Act, so the situation may change. For our sector, the key initiative is the Common European Agricultural Data Space, CEADS. The groundwork was done by AgriDataSpace, a preparatory action funded with about two million euros under the Digital Europe Programme, with fifteen partners from ten countries. It mapped over four hundred existing data sharing initiatives across Europe and recommended building on them, with a decentralised approach that puts farmers' interests at the core. Its blueprint was presented in April 2024. A follow-up deployment project, CEADS, is now turning the roadmap into practice. It describes its vision as a federated data space where public and private data support AI and cloud-to-edge innovation while ensuring data sovereignty for all actors in the agri-food ecosystem, with a consortium from fifteen EU countries. An important point is that the data space builds on what already exists. Rather than replacing existing platforms, it aims to connect them through common rules and interfaces. That is why interoperability, FAIR principles and governance, which we covered earlier, are all part of one picture..

[Audio] A data space only works if participants trust that they remain in control of their data. This idea is called data sovereignty, and it is worth understanding how it can be made practical. One of the best-known technical approaches comes from the International Data Spaces Association, with roots in Fraunhofer research in Germany. Its reference architecture is built around the data space connector, a component that acts as a node in the data space and gives participants the level of sovereignty they need. The idea is that a data provider attaches usage restriction policies to the data before making it available. These policies can state, for example, that data may only be used for a certain purpose, or must be deleted after a given time. The architecture also provides identity management, secure communication and certification of participants. Usage control is the distinctive concept: it goes beyond controlling who can access data to defining and, as far as technically possible, enforcing what the recipient does with it. Technology is only half the story. The other half is incentives. Why do people not share data? The project DIVINE, funded by Horizon Europe, starts from the observation that a broad cost-benefit analysis of data sharing had not been published, and that private data holders are therefore hesitant to share with companies, regulators or researchers. Other factors identified include technical interoperability problems, lack of market transparency and unresolved data ownership questions. DIVINE brings together fifteen partners from eight countries and demonstrates the costs and benefits of sharing agricultural data in four pilots, from dairy and pork to olives and almonds, while developing an agri-data ecosystem that includes governance and policy work. A second Horizon Europe project, CODECS, looks at the same challenge from the perspective of digital ecosystems. It is coordinated by the University of Pisa, brings together thirty-three partners, and works with twenty living labs across Europe. CODECS argues that digitalisation should be understood at two levels: the business processes of individual farms and companies, and the wider digital ecosystems, the networks of organisations, infrastructure and resources that support digitalisation at the farm level. It also stresses that assessments should go beyond economics to include social and environmental dimensions, and that policies must be tailored to each level. Both projects point to the same conclusion. Sharing will happen at scale only when participants can see a fair balance of costs and benefits, when they have technical and legal guarantees of control, and when there are trusted intermediaries to help. For you as educators and advisors, this opens a role: helping farmers and small businesses understand not just whether to share data, but what they would get in return, and under what conditions..

[Audio] Let us consolidate what we covered in this module. We started with data as the connecting thread of the digital value chain. Data comes from many sources and its value comes largely from combining it across actors, which is why questions of who generates it, who holds it, who can use it and who benefits are central. We then discussed ownership, and saw that for non-personal farm data the practical issue is control, access and trust. Farmers have legitimate concerns about reuse of their data, about their weak position in contracts, and about not sharing in the benefits. Transparency and benefit-sharing are key to building trust. On governance, Europe uses both a voluntary and a legislative route. The 2018 EU Code of Conduct on agricultural data sharing set common principles by contract, while the Data Act now gives users binding access rights to the data generated by connected products, with further obligations phasing in. Rights on paper need awareness, tools and support to become real. We introduced the FAIR principles, which make data findable, accessible, interoperable and reusable without requiring that it be open, and we looked at interoperability on four layers, legal, organisational, semantic and technical, with the ISOBUS story showing that standards need testing and governance to deliver. We examined platforms and software ecosystems, which create value through network effects but also create risks of lock-in and concentration of power. And we ended with European data spaces and data sovereignty: the Common European Agricultural Data Space, the Data Governance Act, usage control technologies, and projects like DIVINE and CODECS that investigate how to make sharing worthwhile. The overarching message is that data sharing is not primarily a technical problem. It is a combination of trust, incentives, rules and technology, and no single actor can solve it alone. That is the definition of an ecosystem problem. Now for the reflection. Think about one data flow in your product chain that is not happening today but would create value if it did. It might be quality data from the processor back to the farmer, or field data from the farmer to the advisor. Who holds that data? What would make them willing to share it? What guarantees would they need? If you are watching live, please share your example in the chat. In the next module, we turn to the business side: how do actors in these ecosystems make money, what roles do they play, and what business models are emerging in digital agriculture?.

References. Agricultural Industry Electronics Foundation (AEF). ISOBUS and AEF conformance testing. https://www.aef-online.org/aef-tour/ AgriDataSpace (2024). Blueprint proposal for the Common European Agricultural Data Space. European Commission, Digital Europe Programme. https://digital-strategy.ec.europa.eu/en/library/blueprint-proposal-common-european-agricultural-data-space CEADS project. Common European Agricultural Data Space. https://ceads.eu/ CEMA, Copa-Cogeca et al. (2018). EU Code of conduct on agricultural data sharing by contractual agreement (summary leaflet). https://www.cema-agri.org/images/publications/brochures/EU_Code_of_conduct_leaflet.pdf CODECS project. Maximising the CO-benefits of agricultural digitalisation through conducive digital ecosystems (Horizon Europe, GA 101060179). https://www.horizoncodecs.eu/ DIVINE project. Demonstrating value of agri data sharing for boosting data economy in agriculture (Horizon Europe, GA 101060884). https://divine-project.eu/ European Commission (2017). European Interoperability Framework: Implementation Strategy. COM(2017) 134 final. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=COM:2017:134:FIN European Commission (2020). A European strategy for data. COM(2020) 66 final. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A52020DC0066 European Parliament and Council (2022). Regulation (EU) 2022/868 on European data governance (Data Governance Act). https://eur-lex.europa.eu/eli/reg/2022/868/oj/eng European Parliament and Council (2023). Regulation (EU) 2023/2854 on harmonised rules on fair access to and use of data (Data Act). https://eur-lex.europa.eu/eli/reg/2023/2854/oj FAO. AGROVOC multilingual thesaurus. https://www.fao.org/agrovoc FAO (2022). The State of Food and Agriculture 2022 (Leveraging agricultural automation for transforming agrifood systems). Rome: FAO. https://fao.org/3/cb9479en/online/sofa-2022/notes.html International Data Spaces Association (IDSA). Data Space Connector Report. https://internationaldataspaces.org/idsa-data-space-connector-report/ Jakku, E., Taylor, B., Fleming, A., Mason, C., Fielke, S., Sounness, C. & Thorburn, P. (2019). "If they don't tell us what they do with it, why would we trust them?" Trust, transparency and benefit-sharing in Smart Farming. NJAS: Wageningen Journal of Life Sciences, 90-91. https://doi.org/10.1016/j.njas.2018.11.002 Kruize, J.W., Wolfert, J., Scholten, H., Verdouw, C.N., Kassahun, A. & Beulens, A.J.M. (2016). A reference architecture for Farm Software Ecosystems. Computers and Electronics in Agriculture, 125, 12-28. https://doi.org/10.1016/j.compag.2016.04.011 Rochet, J.-C. & Tirole, J. (2003). Platform competition in two-sided markets. Journal of the European Economic Association, 1(4), 990-1029. https://doi.org/10.1162/154247603322493212 Wilkinson, M.D. et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. https://doi.org/10.1038/sdata.2016.18 Wiseman, L., Sanderson, J., Zhang, A. & Jakku, E. (2019). Farmers and their data: an examination of farmers' reluctance to share their data through the lens of the laws impacting smart farming. NJAS: Wageningen Journal of Life Sciences, 90-91, 100301. https://doi.org/10.1016/j.njas.2019.04.007 Wolfert, S., Ge, L., Verdouw, C. & Bogaardt, M.-J. (2017). Big Data in Smart Farming: a review. Agricultural Systems, 153, 69-80. https://doi.org/10.1016/j.agsy.2017.01.023.

[Audio] Thank you!. thank you!. TALLHEDA has received funding from the European Union's Horizon Europe research and innovation programme under Grant Agreement No. 101136578. Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency (REA). Neither the European Union nor the granting authority can be held responsible for them..