Autonomous navigation is a key skill for assistive and service robots. To be\nsuccessful, robots have to minimise the disruption caused to humans while\nmoving. This implies predicting how people will move and complying with social\nconventions. Avoiding disrupting personal spaces, people's paths and\ninteractions are examples of these social conventions. This paper leverages\nGraph Neural Networks to model robot disruption considering the movement of the\nhumans and the robot so that the model built can be used by path planning\nalgorithms. Along with the model, this paper presents an evolution of the\ndataset SocNav1 [25] which considers the movement of the robot and the humans,\nand an updated scenario-to-graph transformation which is tested using different\nGraph Neural Network blocks. The model trained achieves close-to-human\nperformance in the dataset. In addition to its accuracy, the main advantage of\nthe approach is its scalability in terms of the number of social factors that\ncan be considered in comparison with handcrafted models. The dataset and the\nmodel are available in a public repository\n(https://github.com/gnns4hri/sngnnv2).\n