The creation of large, diverse, and representative gesture datasets is critical for training effective radar-based Human-Computer Interaction (HCI) systems since nowadays many HCI systems use data-hungry computational approaches. However, the resource-intensive nature of manual data collection and the need for extensive labelling pose significant challenges. This chapter explores the potential of synthetic dataSynthetic data  generation as a solution to expand radar gesture datasets. It examines various techniques for generating synthetic radar data, including simulation of radar signals, data augmentation, and the use of generative models such as Generative Adversarial Networks (GANs)Generative adversarial networks. The chapter also discusses the benefits of synthetic data, such as increased diversity of gestures, controlled variability in environmental conditions, and the ability to generate rare or hard-to-capture gestures. Moreover, it addresses the challenges associated with ensuring the fidelity and realism of synthetic data and integrating it with real-world datasets.