This project aims to develop a new approach of classifying SuperDARN (Super Dual Auroral Radar Network) data using machine learning algorithms. In the past, this data has been classified using a formula based on elevation angle, which is not always reliably available, or using another formula based on doppler velocity and spectral width which is biased to miscategorize low-velocity ionospheric backscatter (IS) as ground scatter (GS). Recently, researchers successfully applied machine learning techniques to this data. These approaches improved on past methods, but they used a very limited set of features and relied on simple machine learning methods that do not easily capture non-linear relationships or subtle probability distributions. This project applies machine learning methods with a focus on using a larger number of well-selected features and using more nuanced algorithms. The resulting Github toolkit provides combinations of DBSCAN and GMM for classifying SuperDARN data, along with various plotting tools. This project will continue after the summer, and I plan to add details on validation and try new algorithms.