TileBars
 
Reference: Marti Hearst, TileBars: Visualization of Term Distribution Information in Full Text Information Access. Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems (CHI), pp. 59-66, Denver, CO, May 1995.
URL: http://www.parc.xerox.com/istl/projects/ia/tb-overview.html
TileBars is based on the proposition that “different distribution of term occurrences have different semantics; that is, imply different things” (Hearst 1995).
TileBars is a graphical interface that represents a query result. Given a set of terms, it shows approximately where the words in the retrieved documents are located. Each bar in TileBars is a representation of a document. The length of the bar is an indication of the size of the document. The bars are partitioned into rows and columns. Each row corresponds to a query term. Each column is a segment of the document, partitioned to show the structure of the document. Each cell is colored by a gray value that indicates the number of instances the search term occurs in that segment. By clicking on a "tile" in the TileBar, a text viewer will open to the portion of the document indicated.
One implementation of TileBars uses a TextTiling algorithm to divide the document into non-overlapping segments. The TextTiling algorithm detects segmentation boundaries by analyzing the term repetition patterns within the text. The segment, subtopic, is detected by co-occurrence of sets of terms, locally. Other elements that form the basis for segmentation can be used as well, for example, pages or paragraphs, or document structure.
Tree-Maps
Reference: Johnson, B. and Shneiderman, B., Treemaps: A Space-Filling Approach to the Visualization of Hierarchal Information Structures. Proceedings of IEEE Information Visualization'91, 275-282 1991 IEEE.
URL: http://www.cs.umd.edu/hcil/treemaps/
The initial motivation for Tree-Maps was to visualize large directory structures on hard disk drives. The main objectives of design are efficient presentation, space utilization, interactivity, comprehension, and esthetics. The main concept is to partition the display area in proportion to some weight at the current level and recursively into child levels. Weight of a node must be greater than or equal to the sum of the weights of its children.
The "slice and dice" algorithm partitions the rectangle by recursively dividing the rectangular area horizontally and then dividing the resulting rectangles vertically.
Given bounded boxes, shown as a tree structure, the area can use addition encoding schemes such as color or texture to display attribute value such as file type. Implemented software includes TreeViz and WinSurfer for visualizing directories.
Value bar
Reference: Richard Chimera, Value bars: an information visualization and navigation tool for multi-attribute listings. CHI '92. Conference proceedings on Human factors in computing systems, pages 293-294. ACM.
Value bars are thin, vertical strips added to a window. Generally, they are placed next to a scrollbar, if one exists. When an item is selected in the list, the associated areas in the value bars are highlighted. When an area of the value bar is clicked, the item in a list is selected, and the screen is scrolled to display that item. The height of each vertical bar is the sum of normalized weights of some attribute of the items in the list. For each item, the weight of the attribute is presented as a partitioned portion of the bar. The partitions are stacked and ordered the same as the items in the list. The value bars provide the overview of the weights of attributes of items in the list. Multiple bars may be used for multiple attributes.
VIBE
Reference: Olsen, K. A., Korfhage, R. R., Sochats, K. M., Spring, M. B., & Williams, J. G. (1993). Visualization of a Document Collection: The VIBE System. Information Processing & Management, 29, 1, 69-81.
Korfhage, R. R, VIBE: Visual Information Browsing Environment. Proceedings of the 18th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, Seattle, 363.
Benford, S., Snowdon, D., Greenhalgh, C., Ingram, R., & Knox, I., VR-VIBE: A Virtual Environment for Co-operative Information Retrieval. Eurographics 95. 1995
URL: http://www2.sis.pitt.edu/~webvibe/WebVibe/page02.html
http://www.lis.pitt.edu/~viri/thePage/viri.html
VIBE is a program originally developed for visually exploring multivariate data. Given reference points, Points Of Interest (POIs), are placed in space. Documents are then placed in space according to the ratio of their similarity to the reference points. Moving reference points in space moves document points. Similarity function is calculated from document and reference points vectors. Generally, any interval or ratio data type can be presented in this manner. A single reference point is useless in this model.
VR-VIBE is implemented with the same principle, but data points are projected into 3D using geometric transformation.
The number of reference points dictates the number of dimensions; if there are more than 3 in a 2D situation, the location of data points may be ambiguous. Ambiguity can be reduced by moving reference points.
Document data icons are displayed as transparent rectangles. The transparent rectangles give an indication of how many documents overlap in space. The size of a rectangle approximates how similar a document is to a reference point. In "star" mode, a selected document shows a star of lines pointing to related POIs. The length of each line is proportional to the relative strength of that POI. In Astro-VIBE, a document is shown as a single pixel or as two by two pixels. Astro-VIBE is able to show hundreds of documents at the same time. Selecting an icon shows a list of document titles. Selecting an item in the list shows the document.
Dynamic definition of POIs is implemented in WebVIBE. Users can place selected POIs freely in space or use automatic arrangement. POIs can be deleted or deactivated.
Share with your friends: |
