AUTHOR=Kauttonen Janne , Khan Umair Ali , Aunimo Lili , Nyqvist Antti , Klemetti Aarne 

TITLE=Topic mining for theses and job ads in ICT sector: can higher education institutes respond to job market demands?

JOURNAL=Frontiers in Education

VOLUME=Volume 9 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/education/articles/10.3389/feduc.2024.1322774

DOI=10.3389/feduc.2024.1322774

ISSN=2504-284X

ABSTRACT=In this article, we tackle the pressing challenge of ensuring that higher education students are being equipped with the skills that are topical and in high demand in today’s job market. Our approach to this challenge involved the development of a computational framework that aligns the skills and topics extracted from students’ final-year thesis projects with those described in actual job advertisements in Finland. To accomplish this, we collected Information and Communications Technology (ICT) -related thesis abstracts from an archive of the Finnish Universities of Applied Sciences and used automated subject indexing to extract the top representative topics from texts. Similarly, we gathered ICT-related job ads from a major Finnish job website portal from which we extracted skills sought by employers. These two public data sets covered 12 years with a total of 18254 theses and 107335 ads. The developed framework included a subject indexing model, dimension reduction, clustering, and correlation analysis to find the optimal clusters for theses and job ads, and to compare the two data sets. Our analysis revealed both similarities and differences between the topics and trends present in the theses and job ads, highlighting areas of overlap and gaps. Our framework not only helps to align the education provided with industry demands but also ensures that higher education institutes can stay up-to-date with the latest skills and knowledge in the field, thereby better-equipping students for success in their careers. This framework is not limited to the ICT sector and can be expanded to other fields, providing a data-informed method for the continuous development of teaching and curricula.