The data science methods are widely used in different areas of nowadays life. This paper is dedicated to forming of new approaches, in particular - development of unique model, to provide scientometric research of abstracts from open source pre-print service to process data on subject domains and directions. The objective of research is to analyze degree of presence and importance of data science in different research fields. A new way of working with the information system of the library of the University of Cornelius - the resource of the pre-prints arXiv is proposed in the work. The authors reviewed the abstract information of the resource, which is the result of the search for relevant publications for the given concept. The main attention of the authors was focused on the distribution of publications in the identified scientific areas and the relevant sub-groups provided by the resource. The result of the work is a visual representation and interpretations of the network of subject areas for the concepts - big data, neural networks, deep learning. Keywords: scientometric, big data, data science, concept, subject domain, network, scientific papers