Pasar al contenido principal

SP1: DATAUSE STABLE METHODOLOGIES TO EVALUATE AND MEASURE QUALITY, INTEROPERABILITY, BLOCKCHAIN AND REUSE OF OPEN DATA IN THE AGRICULTURAL FIELD

PRE2020-092585

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de Promoción del Talento y su Empleabilidad en I+D+i
Subprograma Subprograma Estatal de Formación
Convocatoria Contratos Predoctorales
Año convocatoria 2020
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 5
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

FAIR degree assessment in agriculture datasets using the F-UJI tool

  • Petrosyan, Luiza|||0000-0002-1299-8467
  • Aleixandre-Benavent, Rafael
  • Peset Mancebo, María Fernanda|||0000-0003-3706-6532
  • Ferrer Sapena, Antonia|||0000-0001-6432-917X
  • Valderrama-Zurián, Juan Carlos
  • Sixto-Costoya, Andrea
[EN] For the agricultural scientific community, data sharing is crucial both for the advancement of the discipline and ability to meet global challenges, such as the target no. 2, i.e., ¿Zero Hunger,¿ of the Sustainable Development Goals (SDG 2030). In this context, FAIR (Findable, Accessible, Interoperable and Reusable) principles play an important role, as they guarantee the findability, accessibility, interoperability, and reusability of shared data. To improve the practice of data sharing, institutions, funders, and publishers are increasingly demanding data be shared as well as be of an acceptable level of quality, including compliance with FAIR principles. Therefore, the objective of this work is twofold: first, this research aims to determine the degree of compliance with the FAIR principles exhibited by a number of datasets; and second, it aims to explore useful and valid methodologies and procedures that can be used to perform this evaluation quickly, automatically, and effectively. For this purpose, the Data Citation Index (DCI) was used to obtain many datasets in the field of agriculture, which were further grouped by repositories and evaluated using the automated assessment tool F-UJI provided by the FAIRsFAIR project. The results indicated that the principle that exhibited the highest scores was ¿Findable¿, while ¿Reusable¿ received the lowest scores, as none of the analysed repositories achieved a 50% compliance score in this respect. The datasets published in the Zenodo and Dryad repositories exhibited better overall results in terms of the FAIR principles, and the AG Commons repository was the third best rated repository, representing only one of the first three repositories belonging to the agricultural sector. Regarding the use of F-UJI as an automated assessment tool and DCI as a source for obtaining datasets, we conclude that this methodology is useful, and that although it can be improved, it is easy to use and implement by other scientific groups and agents of interest., The authors would like to thank the grant PID2019-105708RB "Stable methodologies to evaluate and measure quality, interoperability, blockchain, and reuse of open data in the agricultural field: DATAUSE" funded by MCIN/AEI/10.13039/501100011033; the pre-doctoral grant (PRE2020-092585) of LP; and the postdoctoral grant (MS21-020) of ASC.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Tesis doctoral (DoctoralThesis)

A FAIR assessment of agricultural data. Applying an automated tool and machine learning

  • Petrosyan Karapetyan, Luiza
[ES] En la presente tesis doctoral se estudia la calidad, interoperabilidad y reutilización de los datos abiertos agrícolas, proponiendo un marco operativo que combina evaluación cualitativa, medición automatizada y análisis basado en aprendizaje automático. En primer lugar, en la Evaluación cualitativa se realiza un diagnóstico inicial de apertura y potencial de reutilización técnica de datos agrícolas en los portales nacionales de Brasil, Colombia y Vietnam. La evaluación se desarrolló en dos etapas: clasificación cualitativa---basada en 14 tipos de datos agrícolas definidos por GODAN---, y revisión técnica---basada en el análisis de características técnicas de los portales---. En la clasificación cualitativa se identificaron deficiencias críticas en áreas como la elevación y la gestión de plagas y enfermedades en todos los portales. La revisión técnica reveló una considerable asimetría estructural. Estas limitaciones ponen de manifiesto la necesidad de implementar un protocolo de evaluación cuantitativa reproducible. A continuación, la Medición automatizada FAIR midió el grado de cumplimiento FAIR en una amplia colección de datasets}agrícolas. Para ello, se utilizó la herramienta de evaluación automatizada F-UJI con datos extraídos a través del Data Citation Index. Los resultados mostraron que el cumplimiento se ajusta, en términos generales, al orden alfabético de los principios: ''Findable'' y ''Accessible'' alcanzaron la puntuación más alta, mientras que ''Interoperable'' y ''Reusable'' obtuvieron un resultado significativamente más bajo. Ninguno de los repositorios analizados alcanzó más del 50% de cumplimiento del principio ''Reusable''. Finalmente, un Análisis mediante técnicas de aprendizaje automático no supervisado permitió profundizar en el diagnóstico FAIR proporcionando una segmentación interpretable. Se utilizó un sistema híbrido de Mapas Autoorganizados (SOM) para la reducción de la dimensionalidad y la Partición Alrededor de Medoides (PAM) para la agrupación. El análisis dividió los datos en tres clústeres: el clúster 1 se identificó con baja alineación, escasa interoperabilidad y reutilización debido a la fragilidad de los metadatos/licencias, el clúster 2 con un perfil equilibrado, buenos metadatos e interoperabilidad, pero muy débiles declaraciones de licencia que limitan la reutilización y el clúster 3 con una alta alineación, especialmente en interoperabilidad y reutilización, aunque necesita mejoras en la estandarización semántica. En definitiva la tesis muestra una progresión desde el análisis cualitativo subjetivo hasta la evaluación cuantitativa automatizada y la interpretación inteligente de datos, consolidando un enfoque escalable y reproducible para diagnosticar y mejorar la calidad de los datos abiertos agrícolas, en línea con los principios FAIR y los objetivos de la ciencia abierta. Los resultados expuestos ponen de manifiesto la necesidad de fortalecer la calidad de los metadatos, las políticas de concesión de licencias y la estandarización semántica para optimizar la interoperabilidad y la reutilización de los datos agrícolas., [EN] This doctoral thesis addresses the quality, interoperability, and reusability of open agricultural data, proposing an operational framework that combines qualitative assessment, automated measurement, and machine learning-based analysis. First, in the Qualitative evaluation, an initial diagnosis is carried out of the openness and technical reusability potential of agricultural data in the national portals of Brazil, Colombia and Vietnam. The assessment was carried out in two stages: a qualitative classification---based on 14 types of agricultural data defined by GODAN---and, a technical review---based on key technical indicators of the portals---. The qualitative classification identified critical deficiencies in areas such as elevation, as well as pest and disease management, across all portals. The technical review revealed considerable structural asymmetry. These limitations highlight the urgent need to implement a reproducible quantitative assessment protocol. Next, the FAIR automated measurement assessed the degree of FAIR compliance in a large collection of agricultural datasets. To do this, the automated evaluation tool F-UJI was used with data extracted through the Data Citation Index. The results demonstrate that compliance generally followed the alphabetical order of the principles, with the ''Findable'' and ''Accessible'' principles achieving the highest scores, while the ''Interoperable'' and ''Reusable'' principles obtained significantly lower results. None of the repositories analyzed achieved more than 50% compliance with the ''Reusable'' principle. Finally, an Analysis using unsupervised machine learning techniques made it possible to deepen the FAIR diagnosis by providing an interpretable segmentation. A hybrid system was used, combining Self-Organizing Maps (SOM) for dimensionality reduction and Partitioning Around Medoids (PAM) for clustering. The analysis divided the data into three clusters: cluster 1 was identified with low alignment, limited interoperability and reusability due to fragile metadata/licensing; cluster 2 with a balanced profile, good metadata and interoperability, but very weak licence statements that constrain reusability; and cluster 3 with high alignment, especially in terms of interoperability and reusability, although it still requires improvements in semantic standardization. In summary, the thesis demonstrates a progression from subjective qualitative analysis to automated quantitative evaluation and intelligent data interpretation, consolidating a scalable and reproducible approach to diagnosing and improving the quality of open agricultural data, in line with FAIR principles and open science objectives. The results presented highlight the need to strengthen metadata quality, licensing policies, and semantic standardization to optimize interoperability and reuse of agricultural data., [CA] La present tesi doctoral aborda la qualitat, interoperabilitat i reutilització de les dades agrícoles obertes, proposant un marc operatiu que combina avaluació qualitativa, mesurament automatitzat i anàlisi basada en aprenentatge automàtic. En primer lloc, a l'Avaluació qualitativa es realitza un diagnòstic inicial d'obertura i potencial de reutilització tècnica de dades agrícoles en portals nacionals del Brasil, Colòmbia i Vietnam. L'avaluació es va desenvolupar en dos etapes: classificació qualitativa---basada en 14 tipus de dades agrícoles definides per GODAN---i revisió tècnica---basada en indicadors clau tècnics dels portals---. En la classificació qualitativa es van identificar deficiències crítiques en àrees com l'elevació i la gestió de plagues i malalties en tots els portals. La revisió tècnica va revelar una considerable asimetria estructural. Aquestes limitacions posen de relleu la necessitat d'implementar un protocol d'avaluació quantitativa reproduïble. A continuació, el Mesurament automatitzat FAIR va mesurar el grau de compliment FAIR en una gran col·lecció de datasets agrícoles. Per a això, es va utilitzar la ferramenta d'avaluació automatitzada F-UJI amb dades extretes a través del Data Citation Index. Els resultats mostren que el compliment s'ajustava, en termes generals, a l'orde alfabètic dels principis: ''Findable'' i ''Accessible'' van aconseguir la puntuació més alta, mentres que ''Interoperable'' i ''Reusable'' van obtindre un resultat significativament més baix. Cap dels repositoris analitzats va aconseguir més del 50% de compliment del principi ''Reusable''. Finalment, un Anàlisi d'aprenentatge automàtic no supervisat} va permetre aprofundir en el diagnòstic FAIR i proporcionar una segmentació interpretable. Es va utilitzar un sistema híbrid de Mapes autoorganitzats (SOM) per a la reducció de la dimensionalitat i la Partició al Voltant de Medoides (PAM) per a l'agrupació. L'anàlisi va segmentar les dades en tres clústers: el clúster 1 es va identificar amb una baixa alineació, molt poca interoperabilitat i reutilització degut a la fragilitat de les metadades/llicències, el clúster 2 amb un perfil mes equilibrat, amb bones metadades i interoperabilitat, però poques declaracions de llicència que limiten la reutilització i el clúster 3 amb una alta alineació, especialment en interoperabilitat i reutilització, encara que necessita millores en l'estandarització semàntica. En resum, la tesi mostra una progressió des de l'anàlisi qualitativa subjectiva fins a l'avaluació quantitativa automatitzada i la interpretació intel·ligent de dades, consolidant un enfocament escalable i reproduïble per a diagnosticar i millorar la qualitat de les dades obertes agrícoles, en línia amb els principis FAIR i els objectius de la ciència oberta. Els resultats exposats posen de manifest la necessitat d'enfortir la qualitat de les metadades, les polítiques de concessió de llicències i l'estandardització semàntica per a optimitzar la interoperabilitat i la reutilització de les dades agrícoles., Este trabajo ha sido financiado mediante una ayuda predoctoral de la Agencia Estatal de Investigación. SP1: Datause Stable Methodologies to Evaluate and Measure Quality, Interoperability, Blockchain, and Reuse of Open Data in the Agricultural Field (ref: PRE2020-092585).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

The compliance to FAIR principles of shared data in addiction research

  • Sixto-Costoya, Andrea
  • Valderrama-Zurian, Juan Carlos
  • Ferrer Sapena, Antonia|||0000-0001-6432-917X
  • Aleixandre-Benavent, Rafael
  • Peset Mancebo, María Fernanda|||0000-0003-3706-6532
  • Petrosyan, Luiza|||0000-0002-1299-8467
[EN] The aim of this study is to assess the scientific data sharing in the field of addictions by applying FAIR principles. These principles play an important role, as they guarantee a minimum of findability, accessibility, interoperability and reusability of the shared data. They are one of the main measures to improve the integrity and quality of research data. For this study, three automated tools were used: the Data Citation Index (DCI) to capture datasets on addictions; Bibliometricos, proprietary software for data retrieval; and the F-UJI tool for the FAIR evaluation of datasets. The datasets on the most common addiction topics, such as alcohol, cannabis, tobacco, cocaine, opioids and stimulants, were downloaded by the DCI (5967 DOIs) and parsed into a database for subsequent analysis. In terms of datasets characteristics, alcohol, tobacco and opioids were the most productive. After assessment by F-UJI, none of the addictions analyzed reached an average of 30% FAIR compliance since all of them were between 20% and 29%. When analyzing each principle, Findable was the best scored principle (in a range of 40%-59%), followed by Accessible, Interoperable and Reusable. The results of our study show, first, an increasing number of shared datasets over the years, especially from basic studies. In terms of quality, there are issues that remain to be resolved, especially in relation to interoperability and reusability principles. This emphasizes the important role of adequate data sharing procedures in ensuring that datasets are FAIR compliant and usable in addiction research., The authors would like to thank the grant "Stable methodologies to evaluate and measure quality, interoperability, blockchain and reuse of open data in the agricultural field" funded by MCIN/AEI/https://doi.org/10.13039/501100011033; the predoctoral grant (PRE2020-092585) of LP; the postdoctoral grant (MS21-020) of ASC; and the funding from the Generalitat Valenciana (Spain) through the PROMETEO 2024 CIPROM/2023/32 grant. Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

FAIR degree assessment in agriculture datasets using the F-UJI tool

  • Petrosyan, Luiza
  • Aleixandre-Benavent, Rafael
  • Peset, Fernanda
  • Valderrama-Zurián, Juan Carlos
  • Ferrer-Sapena, Antonia
  • Sixto-Costoya, A.
For the agricultural scientific community, data sharing is crucial both for the advancement of the discipline and ability to meet global challenges, such as the target no. 2, i.e., “Zero Hunger,” of the Sustainable Development Goals (SDG 2030). In this context, FAIR (Findable, Accessible, Interoperable and Reusable) principles play an important role, as they guarantee the findability, accessibility, interoperability, and reusability of shared data. To improve the practice of data sharing, institutions, funders, and publishers are increasingly demanding data be shared as well as be of an acceptable level of quality, including compliance with FAIR principles. Therefore, the objective of this work is twofold: first, this research aims to determine the degree of compliance with the FAIR principles exhibited by a number of datasets; and second, it aims to explore useful and valid methodologies and procedures that can be used to perform this evaluation quickly, automatically, and effectively. For this purpose, the Data Citation Index (DCI) was used to obtain many datasets in the field of agriculture, which were further grouped by repositories and evaluated using the automated assessment tool F-UJI provided by the FAIRsFAIR project. The results indicated that the principle that exhibited the highest scores was “Findable”, while “Reusable” received the lowest scores, as none of the analysed repositories achieved a 50% compliance score in this respect. The datasets published in the Zenodo and Dryad repositories exhibited better overall results in terms of the FAIR principles, and the AG Commons repository was the third best rated repository, representing only one of the first three repositories belonging to the agricultural sector. Regarding the use of F-UJI as an automated assessment tool and DCI as a source for obtaining datasets, we conclude that this methodology is useful, and that although it can be improved, it is easy to use and implement by other scientific groups and agents of interest., The authors would like to thank the grant PID2019-105708RB “Stable methodologies to evaluate and measure quality, interoperability, blockchain, and reuse of open data in the agricultural field: DATAUSE” funded by MCIN/AEI/10.13039/501100011033; the predoctoral grant (PRE2020-092585) of LP; and the postdoctoral grant (MS21-020) of ASC.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

The compliance to FAIR principles of shared data in addiction research

  • Sixto-Costoya, A.
  • Ferrer-Sapena, Antonia
  • Aleixandre-Benavent, Rafael
  • Peset, Fernanda
  • Valderrama-Zurián, Juan Carlos
  • Petrosyan, Luiza
The aim of this study is to assess the scientific data sharing in the field of addictions by applying FAIR principles. These principles play an important role, as they guarantee a minimum of findability, accessibility, interoperability and reusability of the shared data. They are one of the main measures to improve the integrity and quality of research data. For this study, three automated tools were used: the Data Citation Index (DCI) to capture datasets on addictions; Bibliometricos, proprietary software for data retrieval; and the F-UJI tool for the FAIR evaluation of datasets. The datasets on the most common addiction topics, such as alcohol, cannabis, tobacco, cocaine, opioids and stimulants, were downloaded by the DCI (5967 DOIs) and parsed into a database for subsequent analysis. In terms of datasets characteristics, alcohol, tobacco and opioids were the most productive. After assessment by F-UJI, none of the addictions analyzed reached an average of 30% FAIR compliance since all of them were between 20% and 29%. When analyzing each principle, Findable was the best scored principle (in a range of 40%–59%), followed by Accessible, Interoperable and Reusable. The results of our study show, first, an increasing number of shared datasets over the years, especially from basic studies. In terms of quality, there are issues that remain to be resolved, especially in relation to interoperability and reusability principles. This emphasizes the important role of adequate data sharing procedures in ensuring that datasets are FAIR compliant and usable in addiction research., The authors would like to thank the grant “Stable methodologies to evaluate and measure quality, interoperability, blockchain and reuse of open data in the agricultural field” funded by MCIN/AEI/https://doi.org/10.13039/501100011033; the predoctoral grant (PRE2020-092585) of LP; the postdoctoral grant (MS21-020) of ASC; and the funding from the Generalitat Valenciana (Spain) through the PROMETEO 2024 CIPROM/2023/32 grant., Peer reviewed