Analysis of Users’ Sentiments in Social Media (on the Example of the Astrakhan Region)

Abstract

The article is devoted to the studying of the opinions and sentiments of users of regional communities in the social network VKontakte using methods of machine analysis of text data, supplemented by sociological research methods. In the course of the study, we identified a list of current topics discussed by the inhabitants of the region, determined the most frequently mentioned persons, and analyzed the tone of their mention. Additionally, on the basis of the obtained results, the index of subjective (non-) well-being (ISW) was calculated for each district of the region and a map of the emotional coloring of posts from the communities of the analyzed social network was built. The results of the study can be used to monitor the situation in the region, finding problem areas, elicitation opinion leaders (popular personalities of the region that have a special influence on the opinion of the population), as well as identify the most interesting topics and urgent problems for the population. In perspective, this method of monitoring the social sentiments of the population of the region can be improved by automating the addition of new data to the analytical project. In the future, the addition of mathematical models to the system will make it possible to create graphs for predicting further changes in the region.