Currently, modern social technologies are used by hundreds of millions of users, are available free of charge, attractive and interesting. The article discusses the possibility of the use of social networks to improve e-learning institution of higher education. Considering the large amount of information disseminated by university students on the social network, it is proposed to use methods of data clustering - k-means (k-means) in the article, to personalize the content of educational materials. The results of the research can be used by teachers and instructors of higher education institutions to improve the content of the e-course and personalize e-learning.