Since Jane Jacobs's seminal insights in the 1960s, one of the most emphasized notions in urban studies is the role of architectural and urban form in the 'vitality' and 'liveability' of cities, understood as sets of social and economic qualities such as people's co-presence in public spaces and diversity in local activities. However, can buildings aff ect their urban surroundings? Would diff erent architectural types have diff erent eff ects on the social life of streets and neighbourhoods? These questions are all the more important once we observe a trend in developing countries and other regions – a form of urban growth shaped by detached vertical buildings and gated communities surrounded by setbacks, replacing traditional buildings and creating fragmented urban fabrics. We develop an approach to recognize empirically the urban eff ects of buildings while controlling for systemic factors such as street network eff ects. We apply this method in a large-scale empirical study with twentyfour areas randomly selected in Rio de Janeiro, Brazil. Statistical results suggest distinct eff ects of building types and their features on pedestrian behaviour and land use diversity, helping answer a question that puzzles the spatial imagination: does architecture ma er to urban vitality?