Modelling the impact of mobility on spreading dynamics

The world population increased from 4 billion people in 1974 to 7.8 billion people in 2020 and continues to grow rapidly. Today, 55% of the world's population live in urban areas, a proportion that is expected to increase to 68% by 2050 according to the United Nations. The migration from rural to urban areas coupled with an increasingly mobile population forms a complex contact network that favours the rapid and large-scale spread of infectious diseases. The recent outbreak of coronavirus disease has demonstrated that physical human interactions and modern movement paradigms are the principle drivers for the rapid spatial spread of infectious diseases. Thus, modelling the impact of mobility is crucial to understand the underlying dynamics of the spreading process and consequently to develop effective containment and control strategies. Spreading processes have been widely studied in various different contexts. These processes mainly include the propagation of diseases, rumours and information across a given population network. However, the complexity of human mobility and the limitations on data sources recording movement create challenges for spread modellers. For instance, a main constraint is the use of data that doesn't accurately capture real physical contacts between individuals. Numerous studies considered grouping people based on one of their mobility aspects to identify influential behaviours responsible for spreading diseases through a network. A prominent example is the higher likelihood of individuals who visit new locations to drive the spread compared to those who visit the same locations. However, addressing this topic while considering a single mobility aspect is not sufficient. In fact, there is a great dependency of how far people move or how much they come in contact with others during their travels. So far, previous work has not incorporated multiple mobility aspects simultaneously. We propose a novel classification technique that divides a population into different mobility groups based on their characteristic travelled distance, destination frequency and number of encounters made. This new approach considers multiple aspects of mobility simultaneously and thus uncovers previously unknown interactions between the various aspects that are relevant to disease spread. Extensive simulations reveal which groups contain the most influential spreaders. In addition, we investigate the transmission flows between the different mobility groups to clarify the different roles they play in the spreading process and to identify dominant spreading paths in the network. Furthermore, we study the change in the impact of the mobility groups on the spread over time and uncover a population influence homogeneity threshold, defined by a percentage of infected population at which the identified mobility groups become equally influential to the spread. Our simulations cover various disease spread scenarios. We model both direct (person to person) and indirect (surface to person) transmission scenarios and varying spread-specific parameters such as the infection probability and the suspension time of pathogens. To demonstrate the conceptual analysis presented in this thesis, we construct a large-scale dynamic human contact network from smart card travel data collected in Sydney, Australia. The availability of the recorded spatial and temporal information creates an unprecedented proxy to elicit different travelling behaviours and to study their influence on spreading dynamics. This type of data captures the co-presence of individual in a small space allowing to simulate real and hypothetical disease outbreaks. The work presented in this thesis demonstrates that only through studying various aspects of human mobility simultaneously we can uncover the complex underlying interplay that allows a disease to propagate through a population. Our results can guide concerned authorities and help in the development and implementation of effective disease containment strategies given specific spread characteristics