Currently, there are no all-inclusive methods for visual analysis of ensemble vector fields (EVF) that provide identification of flow trends and general flow similarity over the full extent of transport across ensemble members. Finite-time Variance Analysis (FTVA) provides flow structure information only on particle distributions at the termination of streamline integration. In this paper, we first present a flow structure based on streamline clustering. Second, we discuss a method using streamline clustering to provide information of flow coherence at corresponding spatial regions in the EVF. We consider the regions where bifurcation in flow trends among the EVF members occur. We will also discuss how both methods can be used as a sequential framework for EVF analysis, by using the results of the scalar flow structure to find regions of member flow dissimilarity for further analysis.
Posted by: Nathan Galli
Over the last 300 years, data visualizations in the form of charts and
graphs have become the primary means for communicating quantitative
information. They are pervasive in scientific papers, textbooks,
economic reports, news articles and webpages. In some cases these
visualizations are the only publicly available representation of the
underlying data. Yet, while people can easily interpret data from
charts and graphs, machines cannot directly access it. The lack of
machine readability significantly hinders analysis, reuse and
indexing. Today, a vast trove of information is locked inside data
visualizations.
In this talk I'll present recent tools and techniques we have been
developing to algorithmically read visualizations and extract useful
information from them. I'll then show how we can use the extracted
information to redesign ineffective visualizations and transfer style
between visualizations. I'll show how this approach allows us to add
new forms of interactivity to static charts and graphs, and thereby
bring such visualizations to life.
Posted by: Deb Zemek