Understanding complex urban interactions: representing, visualizing & modelling network flow

Traditional methods, such as Geographic Information Systems, have failed to the capture network flows manifest in Big Data. Addressing this foundational scientific and policy demand, the project will seek to develop methodologies to effectively depict and model such dynamic spatio-temporal network flows.

Aims and objectives

As complex spatial and social systems, cities are characterized by diverse high-density and dynamic interactions that occur at a variety of spatial and temporal scales (Batty, 2005). Delineating and then interpreting such interactions is a fundamental prerequisite of the endeavour to predict and simulate urban change in support of effective and efficient policy making. Flows of materials, people or information, resulting from complex urban interactions, have been extensively researched across multiple disciplines such as urban studies, geography, criminology, planning, transport and computing science. However, with the increasing availability of fine-grained spatial and temporal network flow or big data (from smart cards, social media and environmental sensors), traditional methods such as geographic information systems (GIS) have failed to effectively depict and model the dynamic spatio-temporal network flows embedded in big data. A key reason for this is that current methods of modelling are unable to capture the network flows between entities (i.e., origin and destination), hampering assessment of spatio-temporal dependence and heterogeneity (e.g., Kordi and Fotheringham, 2016). It is in this context that there is a pressing demand for developing new approaches to represent, visualise and model network flow data in urban and regional systems.

The research questions:

To understand the complicated spatio-temporal interactions shaping the patterns, processes and dynamics of network flows, it is necessary to address the following questions:

• What ontological and epistemological definitions can be applied to network flow data?
• How can the spatial and temporal relationships between flows be innovatively represented and visualised in GIS?
• What spatial, temporal and / or spatio-temporal analyses can be developed based on these new representation and visualisation models?
• How can these analytics be used to model dynamic network flows?

The proposed project aims to develop innovative methodologies of representing, visualising and modelling network flows. The specific research objectives are as follows:
1.Review the current literature spanning the definitions, representation, analysis and modelling of network flow data / big data;
2.Design a new method of representing and visualising the spatial, temporal and spatio-temporal relationships between flows in GIS;
3.Develop new flow-based spatial, temporal and spatio-temporal analytics using these methods;
4.Apply these new analytics to model the patterns, processes and dynamics of diverse network flows using big data;
5.Evaluate the roles of these new methods in modelling (e.g., predicting) flows and understanding complex spatial interactions.