This project aims to build a comprehensive visual analytic framework for analyzing large-scale scientific ensemble data. To close the loop between simulation, analysis, and uncertainty quantification, we will investigate automating the workflow in a science application that computes and analyzes an ensemble of models of superconductivity.

We wish to understand how automating the parameter search affects convergence, the number of iterations, and the cost of each iteration to search the parameter space with in situ steering versus a blind search with post hoc processing.

We will experiment with different visual analytic displays of feature-related events and will evaluate usability by our scientist collaborators.