The report describes the experience of using classical machine learning methods to compare the emotional state of individuals living in a metropolis with their sense of self. The basic hypothesis is that each user of a social network leaves a digital footprint; in particular, it is quite easy to determine the locations with which a person most often correlates with it. In the case of a metropolis, such locations are districts. It is obvious that by highlighting the areas with which the individual relates, his comments posted in the public domain can be used to analyze the «emotionality» of the location. At the same time, there are classical sociological practices that can be used to find out the self-awareness of the inhabitants of the districts within the chosen location. So, the result of this study is a comparison of the nature of emotions associated with the area, which can be detected by collecting text data in social networks, and the subjective feelings of the residents of the area when answering questions about the psychological comfort of life in a given location. The practical result of the study is two labeled maps.