By means of Swiss tweet–reply pairs recorded using the Twitter Application Programming Interface (API), we analyze the question of how geographical proximity plays a role in digital communication. Additionally, we investigate how the patterns of Twitter-based communication relate to the functional qualities of space. The results show a strong bias of tweet–reply relationships toward short distances even if we control for age, urbanization or language. While roughly only half of a percent of all messages would occur within the same municipality in a random setting, this was the case for about one-third of all observed Twitter communications. Moreover, the median for tweet–reply distances within Switzerland should have been about ten times larger than observed, if proximity had played no special role. Finally, the degree of urbanization influences the intensity of Twitter-based communication, even if we control for size. The more urbanized a municipality, the higher the probability that a tweet–reply communication will occur.