The field of graph neural networks and urban mobility is rapidly evolving, with a focus on developing innovative methods for analyzing and predicting complex phenomena. Recent research has explored the use of graph convolutional networks (GCNs) for bundle pricing, virtual nodes for efficient long-range information aggregation, and attention-based aggregation methods for vision graph neural networks. Additionally, there has been a surge in interest in applying graph neural networks to urban mobility problems, such as traffic forecasting and train delay prediction. Noteworthy papers in this area include: Learning to Price Bundles: A GCN Approach for Mixed Bundling, which proposes a GCN-based framework for solving the bundle pricing problem. Virtual Nodes based Heterogeneous Graph Convolutional Neural Network for Efficient Long-Range Information Aggregation, which introduces a virtual nodes-based approach for enhancing information flow in heterogeneous graphs. AttentionViG: Cross-Attention-Based Dynamic Neighbor Aggregation in Vision GNNs, which proposes a cross-attention-based aggregation method for vision graph neural networks. RSTGCN: Railway-centric Spatio-Temporal Graph Convolutional Network for Train Delay Prediction, which develops a railway-centric spatio-temporal graph convolutional network for predicting train delays. Fine-Grained Urban Traffic Forecasting on Metropolis-Scale Road Networks, which provides a comprehensive benchmark for traffic forecasting on large-scale road networks.