Considering users' subjective impressions is a challenging question in architectural design. Answering this question is a humancentred approach which is critical for designing spaces that afford true well-being for their users relating to Sustainable Development Goal 3, or SDG 3:"Good Health and Well-being"). While various methods for evaluating users' subjective impressions exist, their focus is mainly on classifying spaces into pre-determined categories representing the users' impressions. Such classification fails to capture the richness of users' actual interpretation, as reflected in descriptions provided by users when they express freely their impression about architectural space. Aiming to capture this richness, we extend the current state-ofthe-art methods and propose an integrated approach to extract and analyse the users' interpretations while interviewing them observing a Tokyo streetscape. In addition, by comparing their comments and gaze, we expose correlations between abstract descriptions and the architectural space, which are difficult to obtain using existing methods. Our insights are a stepping-stone for enhancing computer-aided architecture by integrating visual impressions into the design processes.