AgrospAI, one of the Data Sharing Initiatives (DSI) participating the CEADS project and developed by the project partner Universitat de Lleida, has reached two important milestones: the publication of its data space architecture and use cases in the Computers and Electronics in Agriculture journal, and the implementation of European metadata profiles for its catalogue, including DCAT-AP, GeoDCAT-AP and High-Value Datasets (HVD).
From a demonstrator to validated approaches
AgrospAI is an agri-food data space demonstrator focused on sovereign data sharing and AI services. Within CEADS, it provides a practical environment for exploring how data spaces can enable agricultural stakeholders to share and use data while maintaining control over sensitive information.
A recent journal paper, “AgrospAI: A data space for sovereign data sharing and on-demand AI services in agriculture”, published in Computers and Electronics in Agriculture, presents the architecture and practical experience of the AgrospAI DSI. The work demonstrates the use of the Compute-to-Data (C2D) paradigm: instead of transferring sensitive farm data to external platforms, algorithms can be brought to the location where the data resides thanks to “data rooms”.
The paper illustrates this approach through practical applications including environmental certification for pig farming, where sustainability and compliance reports can be generated without exposing raw operational data, and on-demand semantic mapping, which allows heterogeneous data sources to be integrated as needed, without requiring costly harmonisation in advance.
Making agricultural data interoperable
A second milestone, developed in the context of CEADS Work Package 2 and Use Case 2, addresses another essential requirement for a European agricultural data space: interoperability and discoverability.
AgrospAI has developed a specialised wrapper capable of mapping assets from the Pontus-X ecosystem, originally described using Decentralised Data Objects (DDO), to the European DCAT-AP and GeoDCAT-AP profiles, including the High-Value Dataset (HVD) profile. The implementation is accompanied by a SHACL validation guide, providing a practical mechanism for checking metadata compliance.
The approach has already been applied to a real-world HVDs: the Spanish and French (thanks to INRAE) Land Parcel Identification Systems (LPIS) datasets (SIGPAC and RPG). This demonstrates how existing agricultural data assets can be exposed using standardised European metadata, making them easier to discover, understand and integrate across data-space ecosystems.
Source: AgrospAI portal, https://portal.agrospai.udl.cat/asset/did:op:4449dd9e69d52ca0cd0e4d0fa70e7a5ce87cafdfe2adabc0cd0822d0236cde21
Contributing practical experience to CEADS
Together, these achievements demonstrate how AgrospAI contributes to CEADS at both architectural and implementation levels: from sovereign data sharing and privacy-preserving processing to semantic interoperability and compliance with European data standards.
More Information
Roberto García
Full Professor and Director of the Computer Science and Digital Design Department
Universitat de Lleida, Spain
