Study of social tension based on electronic social networks big data

Abstract

The article presents a methodology for assessing the level of social tension according to the data of electronic social networks. The calculation of the level of social tension according to the specified method is carried out automatically using software tools and requires the participation of the researcher only at the level of an analytical conclusion. An approach to identifying the dissatisfaction of the population at the level of the simplest actions of social network accounts has been described in detail. The necessity of identifying certain spheres of public life for the identification of population discontent has been substantiated, the indicated spheres have been highlighted and described. An indicator has been developed that makes it possible to calculate the level of social tension within each of the identified spheres of public life, taking into account the level of negative messages (posts and comments) and the discussion of each topic. A generalising indicator of the level of social tension has been developed, combining individual indicators of tension by topic. The calculation of the generalising indicator makes it possible to identify the level of social tension in a region or other territorial entity and track its dynamics in any perspective, including retrospectively.The proposed methodology for assessing social tension based on data from electronic social networks differs significantly from existing sociological and statistical approaches. Its main advantage lies in the minimal time lag between the dynamics of social tension, reflected in the social network, and its identification, which expands the possibilities of a prompt response to the growth of negative moods in society. Another difference of the proposed approach is the possibility of multiple repetitions of calculations with minimal, in contrast to the sociological method, the cost of additional resources for each subsequent iteration of the calculation of the indicator value.