<p>Illegal dumping, commonly referred to as fly-tipping in the UK, is a pressing urban management challenge that compromises both municipal budgets and environmental equity. Traditional solutions are largely reactive and unable to identify future fly-tipping events, as they are based solely on historical fly-tipping logs. To address these limitations, we adapt the existing A3T-GCN architecture to sparse fly-tipping prediction and extend it with multisource urban contextual features. The graph component is designed to represent local spatial dependence in fly-tipping risk, providing an operational modelling analogy to the neighbourhood-level spillover implied by Broken Windows Theory. This adapted framework extends crime-count-based prediction by incorporating temporal dynamics and contextual urban information. In addition to historical incident counts, the model incorporates 557 points of interest and land-use features extracted from OpenStreetMap. We simulate the framework using real-world fly-tipping data from Lancaster, UK, that cover the period from 2019 to 2025. The results show that A3T-GCN achieves a 10.4% reduction in RMSE compared with the XGBoost baseline and, more importantly, achieves the highest Recall@50 among the compared models, providing a more than tenfold improvement in hotspot detection efficiency compared with random patrols. Finally, the study provides exploratory evidence on temporal and spatial associations in fly-tipping risk. Specifically, we identify an exploratory two-week temporal pattern in reported fly-tipping, which may be associated with local waste-collection cycles. Similarly, the spatial comparison shows that predicted hotspots tend to coincide with industrial areas and more deprived neighbourhoods, raising an environmental justice question that requires further socioeconomic validation. Our framework enables city managers to shift from a reactive cleaning approach to an equitable prevention approach.</p>