Biological invasion—the introduction and establishment of exotic species outside its native range—is considered as one of the main cause of biodiversity loss worldwide. Amongst the different contributors of biological invasion, human- mediated dispersal is considered to be a major threat (Clifford, 1959; Lonsdale and Lane, 1994; Mack et al., 2000; Mount and Pickering, 2009; Pickering and Mount, 2010; Seebens, Gastner, and Blasius, 2013). Every year, about five per- cent of the global economy is spent on controlling invasive species (Yemshanov et al., 2009). In order to design cost-effective management plans, predictive tools must be developed to identify pathways of exotic species introduction. However, doing so can be challenging, especially in the case of unintended human-mediated dispersal where exotic species are detected only when they are wide spread. Throughout my thesis, I illustrate how network approaches can be used as a predictive tool when studying and managing human-mediated dispersal at different stages of the invasion process. In Chapter 2, my co-authors and I highlight how network analysis can be used as a complementary tool to traditional risk assessments approaches of human-mediated dispersal. While in Chapter 3 and Chapter 4, my co-authors and I demonstrate the application of network models using New Zealand as our case study. We investigated the role of visitors in dispersing exotic species across the country by using the visitors’ travelling patterns as a proxy of propagule pressure to identify areas at risk of biological invasion. As we had no prior knowledge of the visitors’ travelling behaviour, in Chapter 3 we performed an exploratory analysis on only the visitor-place interactions to identify whether visitors had any ‘typical’ way of travelling across the country using a Mixed Membership Stochastic Block Model. Results from Chapter 3 highlighted the importance of accounting for visitors’ behaviour when assessing human-mediated dispersal of exotic species. As both the visitors’ and places’ attributes are good proxies for describ- ing travelling behaviours, in Chapter 4, my co-authors and I exploited both visitors’ and places’ characteristics to predict interactions in the visitor-place network. We extended the Random Dot Product Graph framework using Neural Networks. Results from Chapter 4 highlight how characteristics of visitors and/or places could be used to predict future travelling patterns of visitors as a means to predict future places at risk. By exploiting different aspects of the visitation data using two different network approaches, I show how they improved our understanding of human-mediated dispersal. As a consequence, results from this thesis highlight the application of network models as predictive tools for biosecurity intervention. From a data science perspective, my thesis highlights how one can exploit different data types using different probabilistic network modelling approaches to make inferences despite having limited information.