The system presented in this article was developed for state structures with responsibilities for ensuring security and stability in the society. The main task of the system is to determine the impact of information stored on various Internet resources on society. It serves to assess the level of tension on the basis of comments of social networks’ users on certain news and posts. The system takes into account the number of news’ reposts and the average emotionality of comments on this news. The methods TF-IDF and word2vec are used for keyword determination from the text and their transformation into numerical values. The level of emotionality is defined by an artificial neural network. To get the resulting estimate, a data aggregation block is developed and used to reduce various independent quantitative and qualitative characteristics into one value.