Mapping the thematic structure of Data Science literature with an embedding strategy

The term Data Science covers a family of methods aiming at extracting useful information from large amounts of complex data for decision-making purposes. However, there is a lack of a shared definition, and the diverse communities of scholars and practitioners involved in this research field influence the different conceptualisations. To overview the different methodological approaches and application domains of Data Science and try to depict its thematic structure, we propose a scoping review of the last 10-year reference literature. In the framework of science mapping analyses, we employed a topic detection strategy relying on word embedding to define and describe the how and the what of Data Science, contributing to the ongoing debate about the nature and the peculiarities of this research field.