This paper discusses the technical details of obtaining and processing data to determine a set of characteristics of texts from social networks, genre preferences in movies and music genres Hoses for students of Kazan Federal University who have different academic performance (successful, average, not-successful).The selection of such characteristics is carried out using machine learning methods (Word2Vec, tSNE).The data obtained is used in the development of a functional psychometric model of cognitive behavioral predictors of an individual’s activity within Facial Soap the framework of their educational activities.
We also developed a web application for visualizing the obtained data using the Flask engine.