Stories and conversations in the smart city

If we think of the smart city as a reading environment, we can use it to change what it means to be a citizen, to improve how public topics are addressed, and to democratize how decisions are made. The starting point is text, supplemented with the various other kinds of data that can be gathered through digital means. In this chapter, we discuss two experimental platforms that take different approaches. First is the Data Stories project, where we have been sequencing text from various dynamic sources through a thematic clustering algorithm (Latent Dirichlet Allocation), feeding those thematic clusters into a narrative generator, then putting those results into a storyboarding system. Using the output, we can examine patterns emerging from a variety of text streams, such as Twitter, Facebook, news feeds, and so on. More importantly, however, we can allow people to manipulate the parameters, so that using the same text stream can produce multiple simultaneous valid outputs, depending on the perspective that the reader wishes to take on the feed. Providing a method for encouraging this kind of interpretive or hermeneutic inquiry is a promising strategy for supporting civil discourse. Our second project, Conversational Modeling, is building on previous research to investigate the various ways in which discussions, which occur sequentially through time, can be profitably modeled as 3-D objects of various kinds. These models can subsequently be used for recollection, communication, and analysis, but they may also have a generative potential. As a means of dealing with the structure and substance of discussions in civil society, we propose that conversational modeling has the potential to radically alter our understanding and practice of citizenship.