U.S. Department of Energy

Pacific Northwest National Laboratory

Trelliscope

Trelliscope is a powerful visualization tool for large data that allows users to rapidly render customized graphics that are both detailed (i.e., drilled-down) and interpretable across large datasets — and then filter and sample the results to focus on examples that share common traits or characteristics (e.g., display only plots with correlation above 95% or display plots where regression model fit was poor). Trelliscope is also very useful for visualizing small datasets as well. It is unique in that it is highly customizable and is based on trellis displays which guide users to focus on relatively small, rational subsets of data. This approach helps data scientists develop machine learning algorithms and predictive analytics that account for the scientific phenomenology displayed in the graphics — an essential requirement in many environments.

 

Additional Staff: 
Ryan Hafen, Landon Sego, Luke Gosink
Principle Investigator: 
| Pacific Northwest National Laboratory