Reflecting the Emotions of the City: Developing a Contextual Sentiment Analysis Method for Social Media Texts

The paper describes the developed method for assessing the sentiment of social media texts, taking into account their contextual content. The authors focus on the challenge of monitoring the social mood of city residents with the goal of evaluating the effectiveness of the city's social infrastructure and adjusting it based on the emotions identified among the residents. The texts from social networks are proposed for this process. The study demonstrates that the thematic modeling combined with sentiment analysis is well-suited for this task, creating an "emotion-social sphere-city district" relationship. However, a significant challenge is to identify emotions accurately, as texts from social media are filled with colloquial language, sarcasm, and cultural biases. This suggests that the context of the resident's statements must be considered in order to correctly assess the tone of their messages. The result of the research is the method and developed formula that reveals emotion based on a number of factors and takes into account the context.