OmniViz Visualizations

Each of the OmniViz® visualizations provides a unique perspective of your data.

Galaxy:

The Galaxy, used for all types of data, is a two-dimensional plot of all records in a Data Set, where proximity is a measure of relatedness. Thus, Galaxy shows how every record is related to every other record. Records are shown as points in clusters that resemble small galaxies. Clusters located near each other have similar record content, and clusters far apart from each other have relatively disparate record content. Icons, called cluster "centroids", mark the geometric center of the cluster. For Views based on numeric Column Sets, a cluster centroid is a small circle; in text-based Views, the icon is shaped like a paper document. In both cases, the relative size of the centroid indicates the number of records in the cluster. For most analysts, the Galaxy serves as a "home base" for data exploration because it displays selected records, group highlights or data value colors, and note flags and provides an orientation to the entire Data Set.

Corscape:

Relevant to numeric, categorical and text data, CorScape is a color-coded compressed "spreadsheet" in which each record is a row and each record attribute is a column. Colors are applied to "cells" according to their data values, so that all the patterns that make up a Data Set can be examined quickly.

As in the Galaxy, records are grouped into clusters. CorScape uses correlation or dendrogram ordering of clusters, and records within clusters, to convey how records and clusters are related. Ordering of columns and manipulation of the color scheme bring out interesting features in the data. Hover over any cell to display its name and value. Select records or clusters by clicking on them; selections will be displayed in other visualizations and tools

Comet:

The CoMet visualization is a means for analyzing associations between terms, categories, clusters, and groups. Examples of categorical data include qualitative descriptors such as high, medium, and low responders; the TMN classification scheme used in diagnosing cancers; chromosomal location of a particular locus or gender. Terms are simply words gathered from text fields in records. The CoMet visualization is a 2D representation of the distribution of categorical data or terms across clusters or sets of records. Columns represent terms or individual values from categorical fields, user-defined groups, or terms. Matrix cells are then color-coded to indicate positive and negative correlations. Similar to CorScape, column ordering and changes to the color scheme emphasize interesting correlations.

ThemeMap:

Exclusive to Views clustered on textual data, the ThemeMap visualization is a 3D map of themes or concepts within a Data Set, organized to correspond with the Galaxy visualization. Peaks are assigned to regions containing documents with similar themes, and the height of the peak is dependent on the strength of the themes and how densely the information is concentrated at that location. Use the graphic terrain of the ThemeMap visualization to identify which strong themes are represented in your Data Set and to gain insight into its content. ThemeMap complements Galaxy by displaying the thematic content of a Data Set in a familiar contour scheme.

Treescape:

The TreeScape provides an overview of the cluster relationships in a Data Set, and the hierarchical nature of the relationships for hierarchically-clustered records. Interactive dendrograms can also be constructed for column relationships. A CorScape-like matrix is displayed in conjunction with the dendrogram(s) to provide context. Users can walk up and down the branches of the tree to gain an understanding of the structure in the clustering. Records, clusters, and whole nodes of the tree can be selected for examination in other tools. In addition, tree nodes can be collapsed individually to allow user control of the tree presentation. Finally, new Galaxy visualizations can be constructed based on fine manipulation of the dendrogram.

Correlation:

The Correlation visualization provides an overview of how the numeric attributes in the Data Set correlate with each other. The visualization provides both an high-level presentation of all correlations as well as a detailed plot of the correlation between any two variables. In addition, a relevance network is provided as an alternative presentation. The visualization provides methods for comparing attributes across the entire Data Set and between groups of records. Multiple correlation metrics are available and results can be exported for use in other software packages