FAIR AND MEANINGFUL: SEMANTICS IN DATA SCIENCE FOR GEOSCIENCES (Invited Presentation)

Vocabularies, ontologies, knowledge graphs, and large language models ¾ semantic technologies have always seen quick response from the geoscience community. In the past decades, geoinformatics researchers have built successful applications that leverage semantic technologies to accelerate data flow and scientific discovery in geosciences. Yet, newcomers might feel overwhelmed by the technical terminologies used in this field of research and get lost about what exactly semantic technology can do for their work. This presentation will use an empirical approach to introduce a list of use cases following the steps in a data science workflow, from data collection, pre-processing, and data exploration to advanced analytics, data products, and result communication. Those use cases will illustrate how different semantic technologies, from simple vocabularies to complicated large language models, are used to improve the efficiency of data science for geosciences. While many of them are related to the FAIR (findable, accessible, interoperable, and reusable) data principles, recent trends also show extended functionality of semantic technologies for the meaningfulness and trustworthiness of scientific workflow and results.