Estimating the pose of animals facilitates the understanding of animal motion and can permit the early diagnosis of lameness, improve animal performance in sports such as horse racing, permit robots to learn from animal motion data and produce realistic animal 3D animations and representations.<br> Human pose estimation deep learning models have been able to achieve high levels of performance due to the huge amount of training data available. Achieving the same results for animal pose estimation is harder due to the lack of animal pose datasets. Traditionally, to solve this problem, thousands of images are manually annotated. However, this is labour intensive and expensive. To address this problem, we introduce SyDog: a synthetic dataset of dogs containing 2-dimensional annotations generated using the game engine, Unity. We demonstrate that pose estimation models trained on SyDog achieve better performance than models trained purely on real data and significantly reduce the need for the labour intensive labelling of images.