Different people and cultures associate different emotional states to different parts and spaces of cities. These vary according to individuals, their cultures and also to the time of day, day of week, season, special occasions and more. Recurring patterns may occur in correspondence of the places in which people work, study, entertain themselves, consume, relate, wait or just take a break. What can we learn from these patterns? Trying to find possible answers to this question passes through the possibility to visualize and represent the configurations of emotional expressions in urban spaces, across time, geography, theme, cultures and other dimensions. We have developed ways in which it is possible to harvest people's geo-located (or geo-locatable) emotional expressions from major social networks and to visualize them according to a variety of different modalities. In this paper we will present a series of these types of visualizations, and the ways in which they can be used to gain better understandings of these emotional patterns as they arise, from points of view which derive from anthropology, urbanism, sociology, politics and also arts and poetics. The paper will focus on the ways in which the data is harvested from different social networks, then categorized and annotated with meta-data describing the emotional states, the languages in which people express themselves, the geographic locations, the themes expressed. A methodology for representing this information across a variety of domains (time, space, emotion, theme) will then be presented in detail. A reflection on possible usage cases for anthropology, urbanism, policy-making, arts and design will end the contribution, as well as the description of series of open issues and the indication of possible next-steps for research.