Exploring the possibility of predicting users’ career guidance preferences based on analysis of community topics and the gender in the online social network users’ profiles

The possibility of using digital traces of online social network users, using community themes as an example, to support decision-making in career guidance diagnostics is investigated. Statistical analysis was performed: descriptive statistics, z-criterion for comparing two groups, and regression analysis. The themes of users' subscriptions to various communities available in the social network as well as the gender of the respondent and the number of friends in the social network indicated in the profile were analysed as digital user traces. The socio-professional orientation of the personality was assessed based on the results of the Holland test (edited by G.V. Rezapkina). The correlation between users' digital traces expressed by the themes of subscriptions, and key indicators of socio-professional orientation reflected in the results of the Holland test was analyzed based on the pilot study conducted through an online social networking application. The statistical analysis confirmed the hypothesis that user interests, in the form of community themes, are related to the results of the Holland test. The hypotheses of existing differences in the groups of men and women in the studied attributes (the results of the Holland test and the leading themes of community subscriptions) were proved. By means of regression analysis among the group of women the correlation was found between the prevalence of the community theme “Education” and the key indicators: A (Artistic), E (Enterprising), I (Intellectual); prevalence of the theme “Lifestyle” and severity of the indicators: C (Conventional), I, A, E; “Mass Media” and indicator C. Among the group of men, a correlation was found between the prevalence of “Sports” subject matter and indicator E. The results of the work expanded the space of potential predictors of users' vocational orientation. A foundation for large-scale research in quantifying and constructing predictive models of key occupational indicators based on users' subscription topics has been obtained. The results are useful in the direction of developing an integrated approach to creating a recommendation system for user career guidance.