In this paper we study subjective perception of the city space represented through geography of emotions. In particular, we analyze the continuity of user experience in the city, considering user emotions as connected states. To do this, we develop a concept of an “emotional street network” and analyze the integrity of human emotional experience through availability and connectivity of its emotional network, as well as its valence. To explore the relevance of the emotional street network concept, we use data from a public participation geoinformational system (PPGIS) Imprecity, where users can leave emoji and comments on their feelings in public spaces of 6 types – joy, anger, sorrow, fear, disgust, and surprise. Dataset consists of more than 2000 emotional marks from 600 unique users in open public spaces of St. Petersburg, Russia. Two networks of positive and negative emotions were built: the locations less than 500 m length from each other (a classic measure for pedestrian accessibility) marked with emoji were connected with Points to path algorithm in QGIS software, afterwards collated with the street-road network. The connectivity of the final networks was calculated through axial connectivity index using QGIS Space Syntax plug-in. Both resulting emotional street networks have hierarchical structure – more connected areas in the city center and less connected in the periphery. The negative emotional network is more dispersed, reflecting a more localized and geographically distanced character of negative emotions, covering more peripheral areas than the positive one.