FOCUS on Agricultural Data Interest Group (IGAD) | RDA11 meeting

On 21 - 23 March, 2018 (Berlin, Germany), the Research Data Alliance Plenary (RDA)11: "From Data to Knowledge" welcomed 80+ (RDA Interest and RDA Working) Groups & BoF breakouts sessions, with 661 participants from 41 countries! The RDA11 plenary marks the 5 years anniversary of the RDA, and Agricultural Data Interest Group (IGAD) is proud to have been there since the beginning, and to continue contributing to a number of RDA’s support projects.

RDA FRAMEWORK: WHAT MAKES IT SO SPECIAL?

During the RDA11, different opportunities and challenges of a global data ecosystem - best practices, standards and interoperable data infrastructures fostering implementation of FAIR data, cross-disciplinary knowledge and data/driven innovation - were discussed. In this context, the IGAD focused (19-23 March 2018) particular attention on the cross-disciplinary issues in the implementation of the FAIR (Findable, Accessible, Interoperable, Re-usable) principles.

WHAT AND WHY FAIR DATA? EASIER SAID THAN IMPLEMENTED?

FAIR PRINCIPLES TO ALL ELEMENTS OF
AGRICULTURAL-RELATED
RESEARCH PROCESSES

Under this theme, the Agricultural Data Interest Group (IGAD) |RDA11 meeting was dedicated to the discussion of What does FAIR mean for a community like agri-food? and What does it do about it?

The IGAD event was attended by 66 participants from all over the world who shared their points on (and spread breakthroughs about) FAIR data use-cases from a lot of experience in working with data. Various presentations and discussions welcomed synergies with other RDA Groups, to accelerate FAIR-data-driven innovation together.

 19 March 2018 : the first IGAD|RDA11 day – was dedicated to The FAIR principles:

  • Institutional Use Cases
  • Semantics and Data Models
  • Implementation in Research Infrastructures

- and had 20 presentations covering the broader spectrum of FAIRness. A number of concrete projects for implementing FAIR principles in various agriculture-related communities were presented.

The first session started with opening remarks from the co-chair of IGAD Imma Subirats (FAO of the United Nations, Italy). She stressed the importance to focus on FAIR principles through the scientific and practical dimensions to all research processes to support reproducibility, credibility and impact in agriculture sciences and related disciplines. 

 Thereafter, the keynote speaker - Michel Dumontier (Maastricht University, Holland) raised awareness about the assessment of trust in data by existing evidence-based FAIR data metrics. The participants had the occasion to listen to this perspective in the speech entitled “FAIR: Establishing of a social and technological infrastructure for data discovery and reuse” [CoreTrustSeal CertificationGoFAIRBioArxiv: the preprint server for biology were presented].

 Chandrashekhar Biradar (ICARDA, Egypt), and Enrico Bonaiuti (ICARDA, Jordan) showcased some ICARDA - CGIAR institutional use cases for FAIR principles, speaking about delivery of FAIR Impact in agricultural and Food Security through Big Data which currently is one of the top priorities at the ICARDA-CGIAR institutional level. In this data-intensive scenario, it is very important to represent and link data through trustworthy open valued vocabularies. An interesting use case about keyword maintenance and intelligence - supported by the FAO AGROVOC multilingual thesaurus – for agricultural network analysis for impacts was presented.

 Afterwards, Paul Kasoma (Youths in Technology and Development Uganda, Africa) gave a detailed overview of “Who and How can provide FAIR answers to farmers' questions”. He stressed the importance of making data-driven operational decisions based upon FAIR data, - to optimize yield, to boost revenue while minimizing risks linked to expenses, chances of crop failure and environment impact. The presenter spoke about FAIR data as the roadmap to better food, while drawing special attention to the importance to “talk to farmers” as they have a bigger say in the implementation of the FAIR principles because at one point they can/will be the re-users.

 Rahman Hamidur (Bangladesh Agricultural Research Council, BARC), elaborated more on how FAIR data principles can be incorporated in Agricultural Research Management Information System (ARMIS) database.

 Leanne Wiseman (Griffith University, Australia) showcased how the FAIR principles are embedded into a National Agricultural Data Governance Framework [see also: Legal Dimensions of Digital Agriculture in Australia Report].

The session was followed by Q&A, recommendations and suggestions basing on the previous presentations. 

Among the speakers of the second session - moderated by Patricia Bertin (EMBRAPA, Brazil) - there were:

 Birte Lindstädt (ZB MED Information Centre for Life Science, Germany) remarked the role of the German National Library for Life Sciences ZB MED in the approach to a FAIR Research Data Infrastructure in Agricultural Sciences.

 Richard Ostler (Rothamsted Institute, UK) presented metadata challenges for the world’s longest running agricultural Broadbalk Long Term Experiment (LTE) established in 1843. LTE’s samples collection include grain, soils and vegetation, in particular, relevant datasets for yields, grain quality and nutrients, soil chemistry, crop diseases & weather. Beyond the experiments original purpose, LTE also supports research data and data generation. 

 Antonio Sánchez-Padial’s (Instituto Nacional de Investigación y Tecnologia Agraria, INIA, Spain) presentation focused on implementation of FAIR Principles in INIA’s data repository (read on: REDIA: Designing a Repository for Agriculture and Food Research Data).

 In moving onwards to the next panel round, Daniel Martini (KTBL, Germany) presented practical considerations and challenges for data providers: a case study regarding agricultural planning data at KTBL (attention was basically drawn to semantics and API service). The speaker stated that while the FAIR principles are more specific regarding e. g. requirements for identifiers and necessary conditions for reusability, they also leave some room to interpretation and allow for a split“ data space in which free and open data can coexist with non-free data sets.

 Ashok Patra (ICAR-Institute of Soil Science, India) presented implementation of FAIR principles in ICAR. Since India counts 142 million ha cultivated with cropping intensity 134 %, today, digging up and processing FAIR data from this rich context is a must for decision-making for the benefit of economy and society.

In the breaks among the sections, participants asked questions to the speakers, and communicated with colleagues about challenges faced in: implementation of FAIR principles to meta(data), application CC0 and CC-BY licenses to data, sustainability of standards including vocabularies, identifiers. 

The third session was introduced by Armando Stellato (University Tor Vergata, Italy) who focused on FAIR principles applied to Semantics and Data Models. This approach includes some important criteria: ID/URI that identifies data models should be accessible at a URL through an open protocol; RDF framework should be behind Linked Data that prescribes the resolution of URIs; metadata should be represented by ID in a formal language/schema and format; properties to link to /extend other properties and to describe their license should be in place.

 Clement Jonquet (LIRMM, France) stressed that FAIR data requires good FAIR ontologies and presented related ongoing activities at LIRMM, including AGROPORTAL providing access to a number of open and interlinked ontologies, vocabularies, thesaurus and terminologies described with relevant metadata to facilitate their identification, selection and reuse. He stressed that there is a need to keep transforming community requirements into recommendations, - also for funders, app developers, semantics service providers. [Take a look at the RDA Vocabulary Semantic Services Interest Group Survey].

 Valeria Pesce (GFAR, Italy) spoke about open data standards on support of agri-food sector including those indexed and linked in GODAN Action VEST / AgroPortal Map of Standards (with  380 value vocabularies). The presenter also illustrated findings of a gap analysis on the openness & FAIRness of weather data standards and how these can be assessed against the FAIR principles and TBL5stars (5 star Open Data). The importance to implement open, shared, easy in use standards was particularly stressed: “the easier the standards are to use, the more they are used”. [Read on : a 4-page GODAN Action briefing 'Data standards: how to increase their use in data aiming for impact'].

 Following this impulse to make progress in “semantics” through FAIR Principles, Tom Baker (Dublin Core initiative) shared his vision about the future of the Global Agricultural Concept Space (GACS) able to support fairly different information management contexts relying on semantics.

 Further on, Karel Charvat (Czech Center for Science and Society) presented on “Publication of agricultural Linked Data based on FOODIE model” [for quick overview: FOODIE: Farm-oriented Open Data in EuropeFOODIE Data Models : rationale for Community for On-Farm Data Sharing].

 The following panel on the current state of MIAPPE was presented by Pawel Krajewski (Institute of Plant Genetics, Polish Academy of Sciences). The speaker stated that this is the right moment to reflect upon the importance of data publication and reuse through a checklist that formalize and document the minimal metadata necessary to ensure long term FAIRness of field or greenhouse and phenotyping datasets [a short overview about MIAPPE].

 The next panel on agricultural modelling was presented by Cynthia Parr (USDA, USA) in her speech “AgMIP Data Interoperability: Lessons learned from systems modelling use cases”.

The fourth session was introduced by  Esther DZALE YEUMO (INRA, France), followed by:

 Medha Devare (CGIAR, France) who shared several perspectives on the CGIAR’s Big Data Platform: from the point of FAIR Data to Capabilities.  

 Moving on, Odile Hologne (INRA, France) talked about eROSA project in her presentation called “Towards an e-infrastructure for Open Science in Agriculture & Food”. e-ROSA has launched an OPEN CALL to collect USE CASE studies on specific OPEN SCIENCE in Agri-food distilling (societal, research, data & ICT) challenges that should be addressed in order to achieve Open Science in the agri-food domain.

 Tilemachos Koliopoulos (University of Western Attica, Greece) provided a deep insight on the “Development of an e-­infrastructure to support Monitoring Design Utilities as tool for Efficiency in Agriculture, Food protection and Risk Minimization”, while Giovanna ZappaMilena StefanovaClaudia Zoani (ENEA-SSPT-BIOAG, Italy) presented METROFOOD-RI: a new research infrastructure for linking metrology and food value chains.

 The audience then had the occasion to listen to Théo-Paul Haezebrouck (ACTA, France) who presented API-AGRO: An open platform to share data and API for agricultural stakeholders [learn more about this platform from this article].

 The last talk of the day was delivered by Vassilis Protonotarios (Neuropublic, Greece) who presented “The DataBio use case: Big Data Management and the FAIR Principles”. 

The following core issues permeated all IGAD presentations and contributions: Challenges specific to semantic interoperability – such as “Interoperable” and “Reusable” - should be addressed in a collaborative way, to overcome science-intrinsic obstacles, such as discipline-specific problems.

For more details on the IGAD|RDA11 presentations, you might be interested in viewing the IGAD|RDA11 Agenda and slides on OKAD: F1000Research.

 20 March 2018 : the second IGAD|RDA11 day was dedicated to IGAD group meetings:

The IGAD Working and Discussion groups discussed their accomplishments so far and what actions should be taken forward to achieve what is planned.

 Some highlights from: “Metrics and Indicators in Agricultural Sciences” (agenda)

The group met to discuss about various linked aspects of responsible use of metrics and indicators in agricultural science. While speaking about "FAIRness" - the degree to which a digital resource is Findable, Accessible, Interoperable, and Reusable – which may be defined by increased adherence to measurable indicators, the group also presented its white paper (in progress) on responsible use of metrics and indicators in agricultural sciences. The group has reached out to the key players in the area such as: INRA through ASIRPA ProjectCIRADCSIC-UPVCWTS in Leiden University, WUR.

 Some highlights from: “Agrisemantics” & “On-Farm Data Sharing”

Besides the outputs of this group, it has also presented a roadmap to facilitate the use of semantics in agriculture and nutrition, i.e. to guide integration of metadata production, format conversion, - to support visualization & human validation of mappings as well as access to existing semantic technologies/methodologies. The group conducted a project to demonstrate the feasibility of generating a dataset in the RFD format following the MIAPPE specification from a rich dataset about winter wheat phenotypic data.  The main challenge expressed is the implementation of FAIR data to make them really interoperable.

There is need to develop good practices that will enable and facilitate the free flow and reuse of agricultural data across locations.  It would be a good practice to put farmers into clusters, each sharing their farming experiences…

 Some highlights from: “Wheat- & Rice Data Interoperability”

Following Wheat Data WG Guidelines (included in the FAIRsharing platform and discussed on F1000Research), the Rice Data Interoperability group is working on a cookbook intended for the Rise data managers community. This cookbook will include guidelines on metadata, vocabularies and ontologies plus a decision tree based on data and metadata description recommendations and file format recommendations. The groups also plans to develop and curate a repository of controlled vocabularies and ontologies compliant with the Linked Data standards. This should be the basis for a prospect on multi-lingual conversion of ontologies. Moreover, it is planned to release a Rice-specific data registry and good practices methods for digitization of Rice Legacy Data. How does concept of ‘FAIR use’ apply to copyright law? Even derived metadata could be subject to copyright or license restrictions. Options: (1) Share URLs only, (2) Contact the copyright holder to request permission (chances are they will allow) to share their content (‘data’). This is a complex issue related to copyright (and probably licensing), and the answer may vary depending on the country of the researcher and the location of their database. To be continued…

 During groups discussion there was an interswitch with "Capacity Development" group 
It was intersting to explore various areas of collaborations and how the IGAD groups can be helped in terms of education and training activities focused on agricultural science needs linked to (open) (research) data management issues. [Read on some experiences in the context of RDA/IGAD and GODAN Action]. The Capacity Development group will empower the existing collaboration with GODAN and GODAN Action. This latter is a project to enable the effective use of open data in tackling the food security and nutrition challenges by building the capacity of potential stakeholders, -  to both understand the potential of research and open data for agriculture and nutrition and to engage with it practically. To assess on the current needs within IGAD in terms of capacity development and advocacy, the group will start with Wheat Data WG.

Lifelong learning around data issues is not just a way to boost our skills but also our data-driven economy. Training helps people develop skills and this would also have a tremendous added value for the productivity and competitiveness of the agricultural-related sectors.

 21 – 23 March, 2018, during the RDA11 joint sessions, different RDA groups reported on their activities and on "Move-Them-Forward".

The IG Agricultural Data (IGAD) permeated with a number presentations and contributions interchanging its achievements with RDA representatives from Africa, Asia, Europe, Australia, US.

In particular, IGAD took part in a RDA BoF meeting : Digital Extension Advisory Services: Towards a Data Driven Disruption, and contributed to the discussions on practical issues around the implementation of the FAIR Principles and challenges in the agricultural domain during a joint working meeting among : (1) IGAD (2) IG Data Discovery Paradigms (3) IG ELIXIR Bridging Force (4) WG FAIRsharing Registry: connecting data policies, standards & databases in life sciences (5) IG Repository Platforms for Research Data

CONCLUSIONS OF THE IGAD RDA11 MEETING

1. We all agree that FAIR Principles are important in the agricultural domain. But … FAIR Principles interpretation & implementation can be problematic, due to: 

  • Persistence of identifiers, e.g. pesticides can be only registered for a certain period of time;
  • There are a relevant number of data standards not conforming the FAIR principles – e. g. they lack a mechanism for globally unique identification, do not use open and free protocols for data exchange or do not provide vocabularies that allow data to be semantically machine-accessible, – thus making it more difficult to achieve FAIR conformance. Mapping layers can help here, but there is some work to be done to convince communities that FAIR conformance is a valuable goal to achieve and to specify and implement these mapping layers;
  • Ethical and legal consideration can impact the level of FAIRness.

2. There is a need to enhance targeted training and advocacy to raise awareness about how FAIR data works and which concrete benefits they can derive from a change in data-driven culture. 

  • What is the nexus between data FAIRness vs data value?
  • What does FAIRness mean for the Agriculture community and what do we do with it?
  • How can we achieve interoperability as transparently as possible for data producers?
  • How to agree on right standards?
  • How to provide machine readable licenses?
  • How to provide tools to assess and improve the quality of the data, eg. metadata collection automation
  • How to facilitate the implementation of existing recommendations?

3. Technical requirements for making data interoperable and reusable should be turned into recommendations and awareness about their existence should be widely raised. 

4. Policy enforcement at institutional level and its taking back to the funders is needed.

5. Socio-cultural barrier to data FAIRification needs to be addressed through more credit, incentives.

IGAD|RDA11 concluded with lots of outputs and actions to do until next time we meet in RDA12 : Digital frontiers of global science, as part of the International Data Week|IDW 2018 : #internationaldataweek#idw2018@resdatall@CODATANews@ICSU_WDS@aosp_africa : to be held on 5-8 November 2018, in Botswana (Africa).

It’s every time great to meet the @resdatall community - putting faces to names of collaborators, meeting new, bright people dedicated to #data#researchdata#DataManagementPlan#DataScience
#Datamanagement#FAIRdata#FAIRprinciples#ResearchDataManagement#RDM,
#openscience#OpenData#datasharing and much more! 

LEARN MORE ABOUT RDA PLENARIES

Many thanks to our hosts, partners and sponsors!

The IGAD co-chairs would like to express their gratitude for the support that IGAD and WGs have received from the Global Open Data for Agriculture and Nutrition (GODAN), the French National Institute for Agricultural Research (INRA) and Food and Agriculture Organization (FAO) of the United Nations to organize IGAD|RDA11 series of events.

To keep up-to-date with AIMS news, please, Sign up for AIMS Newsletter, follow @AIMS_Community on Twitter.

And, thanks again for your interest !