Visualizing Relationships in Streaming Multivariate Data

In streaming acquisitions the data changes over time. ThemeRiver and line charts are common methods to display data over time. However, these methods can only show the values of the variables (or attributes) but not the relationships among them over time. We propose a framework we call StreamVis that can display these types of streaming data relations. It first slices the data stream into different time slices, then it visualizes each slice with a sequence of multivariate 2D data layouts, and finally it flattens this series of displays into a parallel coordinate type display. Our framework is fully interactive and lends itself well to real-time displays.