Investigating the Behaviour of Twitter Users to Construct an Individual-Level Model of Metropolitan Dynamics.

In this paper, consideration is given to the use of new forms of social network data as a means to enrich
our understanding of complex structures and activity patterns in urban areas. Specifically, a sample of
Twitter messages (‘tweets’) in the city of Leeds is assembled from publicly available sources, and
spatial and temporal patterns in these data are demonstrated, with special reference to the geodemographic
profiles of service users. It is argued that classical space-time models of individual
behaviour provide one possible framework for the interpretation of patterns, and the process of
attempting to classify activities is begun with reference to the geographical distribution, timing and,
importantly, the content of messages. Some initial analysis is undertaken to examine emerging
networks of interconnection between users and individual users’ spatio-temporal behaviour. In the
discussion, it is suggested that the integration of this form of social data analysis with existing microscale
representations and multi-agent models of city structure and dynamics will provide fertile
ground for future research.