The goal of this project is to analyze the estimations of the magnetosphere in order to explore and characterize both typical and atypical behavior in the physical system by applying data mining and visualization tools on the predicted electromagnetic density values. Specifically, we used unsupervised learning methods (K-means) and four anomaly detection algorithms for our analysis.
Our major deliverable is this website with an interactive visualization dashboard that helps scientists and students to navigate the magnetosphere activities across time.
By choosing different algorithms and time ranges and clicking and hovering over the grids, users are able to see the magnetosphere at different granularity levels and focus on the parts they are mostly interested in, including clusters, concentric views, individual coordinates, as well as the anomalies.
The team is led by NASA JPL Senior Data Scientist Valentino Constantinou and comprised of four graduate students from MSiA Northwestern University
MSiA alumni
Class 2020
Class 2020
Class 2020
Class 2020
https://sites.northwestern.edu/msia/2020/06/19/nasa-jpl-magnetosphere-project/
