BBN Technologies

Developers' Toolkit for Building Information Visualizations

BBN's KineViz Toolkit is a visualization toolkit that can be used stand-alone or integrated with other tools to develop powerful information visualizations. Although the KineViz Toolkit is useful for generating standard statistical displays, its unique power lies in its ability to use motion to illuminate patterns and connections in very complex data sets where traditional highlighting methods prove inadequate.

Motion-more than simple highlighting or color enhancement-makes it far easier to detect patterns and relationships in cluttered data sets while maintaining a view of how the pattern fits into the larger context. These visual cues are especially valuable when analyzing data where the volume and complexity has surpassed our ability to visualize it coherently. The KineViz Toolkit allows users to encode attributes of data using motion and to use motion to query for and highlight patterns, making it easier to detect patterns.

Simulated Motion for Link Analysis

Many problem domains can be mapped into graphs and displayed as node/link diagrams. Typically, nodes represent entities and links represent relationships between the nodes. Such diagrams are used for computer and communications network analysis and management, social network analysis (including terrorist networks) and many other applications. For complex problems, even the best diagrams have overlapping links and nodes and are hard to read. While static highlighting can help to point out relationships and patterns, putting the person or account of interest in motion is much more effective. Motion makes the relationships jump out without losing the context of the complete data set.

Motion Brushing for Cross-Display Correlation

Many analysis tasks require multiple displays. For example, a map, timeline, and several scatter plots might be used to display information relevant to security events in Afghanistan. While these multiple displays can be helpful, they also fragment information by scattering it across multiple visualizations. Analyzing information in this format is difficult, requiring an analyst to keep track of multiple representations of the same data and infer relationships.

Typically users highlight items of interest in one display, and the objects are colored the same in all displays. But when color is already in use (for example, to indicate attack type), this color brushing hides information. Coordinated motion brushing allows users to brush and highlight objects in one display and see the same data similarly moving in the other displays.

This motion brushing technique helps analysts discover patterns that are visible only by looking at multiple displays simultaneously. Using motion to group data points in one display and highlighting those same points in other displays makes it easier to process a much larger quantity of data while preserving essential context.

Data-Driven Motion for Analysis of High-Dimensional Data

When viewing a complex multi-dimensional signal data set from multiple perspectives, coordinating the displays becomes an even greater challenge. For example, it is extremely difficult to visualize the multiple dimensions of image data such as polarimetric and thermal images and combinations of Synthetic Aperture Radar (SAR) and optical imagery, or the various properties of message traffic between computers under different conditions such as normal load, port scanning, and attack.

To reveal patterns in dense, high-dimensional data such as these, the KineViz Toolkit represents each data point in the image as a small, three-dimensional element, with each dimension of the data determining shape, color, location and movement of the visual element. These motion elements, or moxels, can move in three dimensions as well as rotationally, using motion to encode eight different data attributes, in addition to the data visible with traditional color or shape encoding. This technique is especially valuable when analyzing high-dimensional data such as hyper spectral information, gene expressions, epidemiological data, or financial transactions.

KineViz Toolkit Compatibility and Availability

The KineViz Toolkit runs on standard PCs and graphic cards and is independent of the operating system. Integration into other visualization packages is so seamless that the displays the KineViz Toolkit produces match the look and feel of those produced by the other software, and users experience only a brief learning curve before they can tap in to the additional visualization capabilities that the KineViz Toolkit delivers.

The KineViz Toolkit is available through several licensing options.