Level of Detail Exploration of Electronic Transition Ensembles using Hierarchical Clustering

Signe Sidwall Thygesen, Talha Bin Masood, Mathieu Linares, Vijay Natarajan and Ingrid Hotz.
Computer Graphics Forum (EuroVis 2022), 41(3), 2022, 333-344.

Abstract

We present a pipeline for the interactive visual analysis and exploration of molecular electronic transition ensembles. Each ensemble member describes the change in the charge distribution between two different molecular states, together with a set of physical properties. The main targeted tasks of the proposed pipeline are comparison and characterization of electronic transitions and their correlation to physical properties by combining automatic and interactive visual analysis of the ensemble. A quantitative feature vector characterizing the electron charge transfer is introduced that serves as the basis for hierarchical clustering as well as for the visual representations. The interface for the visual exploration consists of four components. A dendrogram provides an overview of the ensemble. It is augmented with a level of detail visual summary for each cluster. A scatterplot using dimensionality reduction provides a second visualization, highlighting ensemble outliers. Parallel coordinates show the correlation with physical parameters. A spatial representation of selected ensemble members supports an in-depth inspection of transitions in a form that is familiar to chemists. All views are linked and can be used to filter and select ensemble members. The usefulness of the pipeline is shown in three different case studies.

[PDF]