The urban environment is a highly dynamic and complex system. Urban dynamics in this complex system is largely reflected by the movement of people to and from Places of Interest (POIs) in the urban area. To better understand and plan for the city's various scenarios, there is a need to forecast urban dynamic conditions in terms of the possible movements of people across POIs. However, such predictions are not easy because an interdependent and living system is hard to forecast. In addition, the commuting and shopping of individuals in urban environments will show distinct patterns at various stages of disasters as compared to normal situations. This paper presents a network science-based urban forecast dashboard, in order to monitor urban events and identify the interdependencies that characterize urban dynamics. Behind the dashboard is a deep learning model that incorporates the network dynamics between POIs. The dashboard powers the prediction of urban dynamics from a network science perspective. This research calls for a unified framework to model the flow and network in the city. The dashboard visualizes how network science and urban science can mutually benefit from each other.
Keywords: Dashboard; Events; Network Science; Resilience; Urban Forecast.