MVST-SciVis: narrative visualization and analysis of compound events in scientific data

There is a large volume of spatiotemporally correlated multivariate data in multiple layers of the earth’s environmental system. Compound events arise from the interaction of multiple variables. Current approaches employed by earth scientists lack the flexibility to identify the drivers and corresponding impacts of different events. In this paper, we present MVST-SciVis (MultiVariate SpatioTemporal Scientific data Visualization), a new visual analytics prototype to help scientists explore spatiotemporal correlations among multiple variables, and analyze the drivers and influences of different compound events. MVST-SciVis provides coordinated maps, scatterplots, line charts and bar charts to support a three-level multi-granularity complex visual analysis pipeline. MVST-SciVis also provides a storyline visualization tailored for scientific data that abstracts inter-entity relationships and the driving components information of compound events. Our case studies with the data from two ecosystem circles of climate and agriculture illustrate the usefulness and effectiveness of MVST-SciVis. Graphical abstract