2023Annual Report of the Advisory Group on DataEuropean Food Safety Authority (EFSA),Fabrizio Abbinante, Sandra Basic, David Foster, Klemens Fuchs,Nina Georgieva, Ákos Jó？wiak, Christine Nellemann, Andrea Maldonado,Michele de Martino, Eileen O’Dea, Eva Scharfenberg, Martina Stack,Mikko Tuominnen,Jean-Luc Volatier, Koen WienkAbstractThis report of the European Food Safety Authority Advisory Group on Data presents the results of the group’s activities in 2023. Thegroup’smission is to demonstrate how digitalisation can improve food safety and nutrition in Europe. In 2023,the group carried out asurveyamongitsstakeholderson data-related priorities. The group and its subgroupsdeveloped and prioritised new project ideasin the keyareasand monitored the progress of ongoing multi-country data projectsinitiated by the group. Thegroupliaised with the European Commissionon ongoingdata initiativesand discussed the implications and opportunities of digitalisation as well asthe role ofartificial intelligence(AI)for the European food safety data ecosystem.© European Food Safety Authority, 2024Key words:Advisory Group on Data, data, digitalisation, digital ecosystem, interoperabilityRequestor:EFSAQuestion number:EFSA-Q-2023-00751Correspondence:idata@efsa.europa.eu

1IntroductionThe final report(EFSA, 2020)of the Advisory Forum Task Force (TF) on Data Collection and Data Modelling was endorsed in July 2020 at the 76thmeeting of the Advisory Forum (AF) of the European Food Safety Authority. Thatreport was the culmination of a two-year-work by the Task Force,which was chaired by Hungary. The report outlined a vision for a future food safety data system that would enableEuropean cooperationto maximise the use of information technology and analytical capabilities to further improve efficiency and effectiveness in relation to the safety and sustainability of the food chain.The report contained a total of45 recommendations,which were categorised into 25 operational and 20 strategicrecommendations.These were further grouped into short, medium, and long-term timeframes, covering four key areas:data reporting, data modelling, IT architecture and data analysis. The report highlighted the need tofurther exploreand prioritisethe objectives inthese areas. It alsorecommended the establishment of a standing working group on data-related issues tocarry out this work. This recommendation of a standing working group on data was accepted by the AF and an Advisory Group on Data (AGoD) was establishedtoformally begin its work in April 2021. The function of the AGoDis to act as a guardian and governance body for the current recommendations, monitor their implementation and ensureoverall alignment with EFSA’s data roadmap. The AGoD also acts as a thinktankto generate project ideas, monitor and steer such projects and oversee their implementation.The group is composed of representatives of the AF (on a voluntary basis), technical experts delegated by members of the AF, EFSA staff members and a representative from the Directorate-General for Health and Food Safety(DG SANTE). The group is chaired by Hungary. The AGoD was established for an initial term of 3years(April 2021-April 2024). The AF can extend the mandate of the group for a longer period. The expected deliverables of the group are:？Progress updates atAF meetings;？Multi-country projects agreed, launched, and implemented;？Input to EFSA strategic prioritisation processes in the field of data.The group published its first annual report in April 2023.2Advisory Group on Data -Activities 20232.1MeetingsIn line with its terms of reference,the AGoD held seven meetings in 2023, four of which wereheldvirtually and three in person to facilitate the workshops organisedby the group. Each meet-ing was heldin parallelwith the AF meetings to allow forfeedback and discussion with the AF plenary members.While the online meetings focused primarily on the exchange of information,the monitoring of project progress and the coordination of the sub-working groups, the three on-site workshops focused onthe development of new priorities, the evaluation of project ideas Advisory Group on Data2023Annual Report www.efsa.europa.eu/publicationsEFSA Supportingpublication2024:EN-87606developed by the sub-working groupsin the new priority areasand the drafting of terms of referencefor the new mandate.2.2Subgroups of the Advisory Group on DataThe group has six thematic sub-working groups, of which the following fourstarted their work between 2022and 2023, with the mandate todevelop and share:？Tools and technologies;？Innovative data analytics and new data streams;？Data quality;？Data modelling and terminology.The subgroups “Digital platforms and ecosystems” and “Data literacy and data capacity” will, if confirmed,start their work in 2024.2.2.1Subgroup on developing and sharing tools and technologyIn December 2022, a working group focusing on exploringand sharing tools and technologieswas established. This group's primary objective was to investigate the existing tools and tech-nologies used inMember States that could be potentially usedby others. The aim was to identify opportunities for cross-utilisation and sharing of resources to enhance the operational processes related to data reporting to the European Food Safety Authority (EFSA).During2023, this group metsix times in an online format, focusing on identifying the tools and technologies needed to streamline these processes.The year endedwith a productive two-part workshop. The group reached a consensus onareaswherekey tools are needed: data validation, data enrichment, and a tool tobrowselegal limits. Itisworth noting, however, that while several Member States have already developed tools in these areas, their integration into existing systems has made it somewhat challenging to share these resources more widely.Looking ahead to2024, the group intends to broaden its scope of discussion. This will include not only a continued focus on operational tools but also engagement with EFSA on more strategic matters. Topics for discussion includethe desired potential update of the Standard Sample De-scription version 2 (SSD2)framework and the exploration of API services to strengthenauto-mated data enrichment. These are the discussionsenvisagedto contribute positively to the overall data management and reporting processes. 2.2.2Subgroup on innovative data analytics and new data streamsThis subgroup wasestablished to discuss issues related tonovel data analytic approaches in-cluding artificial intelligence (AI) and new data streams including existing data streams outside the current food safety domain. The following topics were considered as priority areas for the subgroup: a) Enhancing services using new data streams, big data and computational analysis; b) Developing and/or implementing new algorithms (prevalently based on AI methods); c) Using AI-based solutions in risk analysis frameworks; d) Trust, legal, ethical and transparency ques-tions of AI-based and other advanced computational solutions; e) Exploringthe use of tracea-bility data and other new data streams (structured and unstructured); and f) New approaches tofood consumption and composition data collection. Advisory Group on Data2023Annual Report www.efsa.europa.eu/publicationsEFSA Supportingpublication2024:EN-87607The subgroup started its activities at the end of 2022,with 14 members from 10 Member States and EFSA,and had a total of9 (virtual) meetings. After definingthe scope and planning the subgroup activities for 2023, the discussions mainly focused on exploring and developing project ideas. During this process over 40 development ideas wereidentified, and by the end of the discussions 5 projects were considered mature enough to be brought forward for AGoD discus-sions (see later in Chapter 2.4). For 2024, a more focused discussion on the use of AI method-ologies is envisaged, in line with thekey priority areas of both the AGoD and the subgroup.2.2.3Subgroup on data qualityThe working group was established in June 2023 to addressoverarching issues relatedto data quality, to addresstwo recommendations from the final report of the Advisory Forum Task Forceon Data Collection and Data Modelling and, due to the ubiquitous existence and processing of the topic of data quality, to focus on documents and supportnotaddressed elsewhere.As part of thiswork, developments at EFSA such as the KPI dashboards will continue to be monitored.The following concrete work has been started: 1) a guideline for checking data quality criteria that can be used to analyse unknown data in order to use these data and 2) a guidance document on how to describedatasets submitted to EFSA, for example:metadata describingthe limita-tions, their selection, possibletest criteria of the reported data, etc.Furthermore, the group works on data quality issues based on specific use cases and then triesto address the recommendations to the AGoD network (such as other sub-working groups, the networks, EFSA). This may also result in new projects.2.2.4Subgroup on data modelling and terminologyThe main objectivesof this subgroupare to develop and promote ontologies to supporta frame-work for food safety data analysis, to foster collaboration on data models and to reporton re-quirements during the developmentof relevant legislation. It alsoaimsto encourage the use of standardised European data catalogues for Member States inEU reporting and data exchange systems,and to identify ways to improve interoperability between databases. During the second meeting,two presentations were given on a“Smart coding”app and other tailor-made activities relatedto data models and terminology: progress to date and ongoing work on terminology at EFSA. After the discussion, the group concluded that catalogues devel-oped by EFSA such as Foodex2 were an opportunity for Member States to apply standardiseddata models. The main improvement identified concernedthe information exchange between EFSA and Member States. Changes to thecatalogues should be avoided during periods of trans-mission of data from Member States to EFSA. New tailor-made activities, especially the devel-opment of a shared mapping too,wereseen asan important way of using EU data models at Member State level. During the third meetingof the subgroup,EFSA presented the “Rebuild project”to improve the DCF system with a specific focus on data models. It was concluded that this project will clearly help Member States to use the EFSA catalogues. During the 4thsubgroup meeting,ANSES(the French Agency for Food, Environmental and Occupational Health & Safety)presented a toolto automatically codify foods into food groups using a machine learning tool. The use of this kindof new toolfor EU data models was therefore discussed