AGROVOC & the MEL platform: Standardizing metadata and following FAIR principles

FAO/Nadine Azzu

A use case by ICARDA-MEL

The International Center for Agricultural Research in the Dry Areas (ICARDA) is an international organization undertaking research for development providing innovative, science-based solutions for countries across the non-tropical dry areas. It is part of CGIAR, the world’s largest publicly funded agricultural research network. Together with other programmes, projects and organizations (PPOs), ICARDA uses the online platform Monitoring, Evaluation and Learning (MEL) to plan, manage, monitor, evaluate, report, and share its activities and results.

The platform, by nesting PPOs under one common framework and providing a collaborative space, helps users collect, visualize, and use data to better inform decision-making and ensure accountability and transparency for research and development investments. The Monitoring, Evaluation and Learning team at ICARDA produced a video to show how such teams use the platform to maximize impact by managing, monitoring, and evaluating performance for better-informed decisions and results. To reach the fulfilment of its mission, to enhance resilient livelihoods in the dry areas, ICARDA is committed to Open Access Knowledge Management, aiming at providing collaborators with access to data assets, tools, protecting Intellectual Property, maximizing access to research, developing knowledge management capacities and institutions, enhancing scientific knowledge, and generating more impact.

Since ICARDA’s inception in 1977, ICARDA and FAO have been collaborating on topics of agriculture, food security, knowledge management, and data exchange. References to AGROVOC terms and concepts have appeared in most of ICARDA’s official reports and this fruitful cooperation expanded, in 2018-2019, with the implementation of MEL-AGROVOC interoperability, in which AGROVOC thesaurus was implemented in MEL for easy access to its vocabulary data.

The use case based on an information system(s) or practical use case(s)
An important component of MEL’s use is the storage of datasets and documents while providing exhaustive metadata for each item. The MEL platform introduced the use of AGROVOC to describe user profiles, blogs, outcome stories, data assets such as journal articles, research reports, and capacity development events, to mention a few. Each author can select pre-existent keywords or generate ad hoc ones, and the use of AGROVOC keywords is highly recommended. Figure 1 illustrates the use of AGROVOC as a ‘Controlled list’ in MEL.

Figure 1: Use of AGROVOC as a ‘Control List’ across different areas. Source: ICARDA, 2021

How are the keywords collected from MEL? 
The keywords are collected from different places in MEL, from reporting a publication to innovations, outcome stories and datasets, amongst others. 

The user has the option to add, see Figure 2:

  • A new keyword which will be saved in the MEL database.
  • A keyword matched with AGROVOC.

A keyword that already exists in the MEL database.

Figure 2: Options available for the MEL user to showcase how keywords are collected. Source: ICARDA, 2021


It is the task of the knowledge and data management teams to review and update the keywords’ section including as many references to AGROVOC as possible. The innovative approach is not only to provide AGROVOC keywords in MEL and related platforms but also to have a real-time connection with AGROVOC following its monthly updates. This enables a continued synchronization between AGROVOC, MEL, and external repositories (e.g., DSPACE, DATAVERSE) and increased interoperability for broader utilization of data across organizations in accordance with the FAIR principles (Findable, Accessible, Interoperable and Reusable). Each month, the MEL platform automatically reviews all keywords not previously recognized by AGROVOC and matches them with the new concepts or terms provided by AGROVOC, see Figure 3.

The Word Clouds show the most used AGROVOC keywords used within MELSpace and MELData, see Figure 4.

Figure 3: Keywords used to publish MEL data assets in (a) (MELSpace) (b) (MELData). Source: ICARDA, 2021

Figure 4: Word Clouds with the most used AGROVOC keywords from MELSpace and  MELData. Source: ICARDA, 2021

All keywords used in MEL are listed in a specific section of the MEL platform, called ‘Keywords Intelligence’, where it is possible to monitor keywords frequency and check their context of use, see Figure 5. The MEL team actively contributes to AGROVOC thesaurus either through direct submission of common and important concepts from scientists or through activities involving the submission of keywords belonging to a particular domain. Currently, the MEL team is working on the submission of concepts and terms related to the livestock field within the CGIAR Research Program on Livestock by choosing keywords found in MEL that are unmatched with AGROVOC.     

Figure 5: MEL’s Keywords Intelligence page. Source: ICARDA, 2021


Benefits of using AGROVOC
Both AGROVOC and MEL work together in accordance with the new CGIAR Open and FAIR Data Assets (OFDA) policy promoting open and FAIR data as possible. As AGROVOC follows the FAIR principles, the use of the thesaurus benefits and strengthens MEL’s alignment with FAIR principles for many different scientific outputs.
AGROVOC, through its Application Programming Interface (API), provides quick and easy access to its data. The use of AGROVOC keywords when publishing datasets helps in standardizing MEL metadata by reducing the use of ad hoc keywords. This has a positive impact from an archival point of view, allowing the identification of datasets related to similar topics. Since MEL datasets often include local terms in different languages, the matching of keywords within datasets with AGROVOC increases the value of data by providing an immediate translation in multiple languages. 

The standardization and the datasets interconnection improve the data findability. Similarly, for journal articles and grey literature, the use of AGROVOC keywords reinforces FAIR principles, in particular the findability of scientific knowledge and data.