|
Abstract:
|
Trelliscope is a system for detailed interactive visualization of potentially very large data sets. The concept is to break a dataset into pieces, make a plot for each piece (each plot called a panel), and then arrange these plots into a grid or pages of grids for viewing. This idea is based on Trellis Display. When the number of panels becomes very large, Trelliscope provides a system for allowing the user to interactively explore the space of the panels based on metrics computed for each panel, called cognostics. In this talk, I will illustrate applications of Trelliscope that highlight the utility of the approach as a general framework for interactive visualization while also highlighting the scalable nature of the approach Demonstrations will also feature recent developments in the design of the Trelliscope viewer and research enabling automatic context-aware cognostic computation.
|
