Charts and graphs provide a way to convey information in a pictorial manner that is usually better understood by most audiences. Sometimes however, these graphs may contain additional colors and styles that do not add useful detail to the information relayed and may sometimes even make the intended data more complicated to assimilate. In other instances, it is necessary to convert these graphs into a style that is better understood by a specific audience. This project explores various machine learning methods to translate 2D plots and bar charts, from normal charts with different styles, to simpler sketch-like representations that are easier to understand. The results obtained at the end of this project show that various deep learning methods like adversarial learning with Generative Adversarial Networks can indeed produce plausible results for image-to-image translation tasks specific to charts and graphs.