IRIT researchers awarded at the MTSR 2021 conference

A group of researchers from IRIT’s MELODI team composed of Amina Annane, Mouna Kamel, Nathalie Aussenac-Gilles, Cassia Trojahn, Catherine Comparot and Christophe Baehr from the Centre National de Recherches Météorologiques of Météo France (CNRM), has been awarded at the Metada and Semantics Research Conference (MTSR) 2021. The 15th Metadata and Semantics Research Conference awarded the 2nd best paper prize to the research work entitled “Towards the FAIRification of Meteorological Data: a Meteorological Semantic Model”.

A semantic model to improve weather data FAIRization

According to OuvrirLaScience.fr, the FAIR principles are a “set of guiding principles for managing research data aimed at making it easy to find, accessible, interoperable, and reusable by humans and machines.” The focus of the award-winning working group is specifically on weather data. The introduction to the article puts into perspective the issues related to this field. Indeed, “making weather data FAIR in order to facilitate its reuse is a strategic issue because it is essential data for scientific research in many fields. This paper proposes a semantic model combining a metadata model and a data model to describe meteorological observation data. Indeed, the modeling of (meta)data is an essential step towards their FAIRization. We use the “SYNOP” dataset from Météo France to illustrate the difficulties related to accessing and understanding this type of data, and to show how the proposed model improves their adherence to the “F”, “I”, and “R” principles.“

A multidisciplinary approach

This research work is part of a multidisciplinary approach. It is indeed deployed in the framework of the ANR Semantic4FAIR project and is supported by the DataNoos alliance. The goal of the Semantics4FAIR project is to allow scientific communities to easily find, adapt and reuse data produced by Météo France. The aim is to provide tools to describe these data in a way that is adapted to future users, whereas current metadata and descriptions are, when they exist, based on the point of view of the data producers.

DataNoos, the transdisciplinary academic alliance on digital resources and knowledge practices, explains the interest of an interdisciplinary approach: “To facilitate access to data by non-specialists in meteorology, we proposed an interdisciplinary response involving computer scientists, researchers in human sciences (ergonomics) from the Maison des Sciences Humaines et Sociales de Toulouse (MSHS-T) and in meteorology, data producers (Météo France) and users of these data (OMP). We have chosen an approach using ontologies and formal vocabularies that uniquely define the concepts, properties and entities necessary to define rich and understandable metadata. The semantic representation allows reasoning about this data when searching it, and facilitates its alignment with other open data.“