InfoSky: Difference between revisions

From InfoVis:Wiki
Jump to navigation Jump to search
Line 82: Line 82:
training on the browser’s features.
training on the browser’s features.
The test dataset consisted of 80000 newspaper articles from the German Sueddeutsche Zeitung. Tasks included locating a document or collection within the hierarchy, counting the number of documents in a collection, comparing the number of items in different collections and counting the number of similar documents in a collection.
The test dataset consisted of 80000 newspaper articles from the German Sueddeutsche Zeitung. Tasks included locating a document or collection within the hierarchy, counting the number of documents in a collection, comparing the number of items in different collections and counting the number of similar documents in a collection.
At the end of each test, an  
At the end of each test, an  
interview was conducted with the test user to gain additional  
interview was conducted with the test user to gain additional  
feedback. The entire session was videotaped.
feedback. The entire session was videotaped.

Revision as of 14:36, 12 April 2007

Authors


Short description

The InfoSky visual explorer is a system enabling users to interactively explore large, hierarchically structured document collections. Similar to a real-world telescope, InfoSky employs a planar graphical representation with variable magnification. Documents of similar content are placed close to each other and displayed as stars, while collections of documents at a particular level in the hierarchy are visualised as bounding polygons. Usability testing of an early prototype implementation of InfoSky revealed several design issues which prevented users from fully exploiting the power of the visual metaphor. Evaluation results have been incorporated into an advanced prototype, and another usability test has been conducted. A comparison of test results demonstrates enhanced system performance and points out promising directions for further work.
[Michael Granitzer et al., 2004]



Navigation through InfoSky (Current Version)

  • Selecting a collection: Left-clicking a collection label selects the collection and auto-centres it.
  • Selecting a document: Left-clicking an individual star selects the corresponding document and auto-centres it.
  • Selecting the parent collection: A toolbar button allows to place the focus on the parent collection. The viewport is zoomed out to display the collection.
  • Continuous hierarchical zoom: After selecting a collection, zooming in on the visualisation continuously selects deeper hierarchical levels based on magnification and position.
  • Panning: Dragging with the left mouse button pans the viewport. Collections are auto selected based on magnification and position.
  • Zooming: Using the mouse-wheel, the magnification factor of the display can be adjusted.

Algorithms used for Implementation

  1. Force-directed Placement Algorithm
  2. Geometric Transformation
  3. Area Partition

Suitable Datatypes

Suitable Datatypes for this information visualization technique are hierarchical information structures. The goal of InfoSky is to reduce the complexity of large amounts of text-data (for examples collections of newspaper articles) by presenting it through the metapher of a telescope view in a hierarchically structured form. Text-data can be either Word documents, PDF or HTML files.


Figures

Pictures of the InfoSky Browser 2002

InfoSky
InfoSky



Reengineered InfoSky Browser 2004

InfoSky
InfoSky



Important Citations

Main challenges of Infosky:

Hierarchy plus similarity: Represent both the hierarchical organisation of documents and inter-document similarity within a single, consistent visualisation.
[Granitzer et al., 2004]



Focus plus context: Integrate both a global and a local view of the information space into one seamless visualisation.
[Granitzer et al., 2004]



Stability: Use a stable metaphor which promotes visual recall and recognition of features. The visualisation should remain largely unchanged at a global level even if changes occur to the underlying document repository on a local level.
[Granitzer et al., 2004]



Unified frame of reference: Support a single, consistent view of the document space for all users, regardless of the access rights of each individual user, thus providing a common frame of reference for all parties.
[Granitzer et al., 2004]



Exploration: Provide simple, intuitive facilities to browse and search the repository. The visualisation tool should allow the visualisation to display a maximum number of document properties and relationships without any need for user interaction. It should thus offer a means of locating documents without specifying a query, by simply browsing the information space and displaying information within its context.
[ Granitzer et al., 2004]



Scalability: Visualise very large (hundreds of thousands, if not millions of entities), hierarchically structured document repositories.
[Granitzer et al., 2004]




Evaluation

Initial Study

Study in 2004

Tested User Interface

The interface of the tested prototype was divided into three distinct areas, the tree view (top-left), the galaxy view (top-right) and the list view (bottom). Combinations of tree and galaxy view were tested in the experiment, with the list view remaining in place.

Changes derived from initial study
  • a list of documents in the selected collection was added
  • layouting of labels in the galaxy view was revised to minimize cluttering
  • interactions were optimized based on user feedback received during the initial study

Test Setup

Nine employees of the Know-Center were recruited for the study and divided randomly into three groups of three. Users of the first group began with the telescope browser (condition GV), then used the tree view (condition TV), and finally the mixed view (condition MV) displaying both the tree and the telescope view . The other two groups used alternating ordering of test conditions. Before using the telescope browser (either stand-alone or in conjunction with the tree view), users were given a brief training on the browser’s features. The test dataset consisted of 80000 newspaper articles from the German Sueddeutsche Zeitung. Tasks included locating a document or collection within the hierarchy, counting the number of documents in a collection, comparing the number of items in different collections and counting the number of similar documents in a collection. At the end of each test, an interview was conducted with the test user to gain additional feedback. The entire session was videotaped.

References

Further Information

  • Andrews, K., 2006. Evaluating information visualisations. In Proceedings of the 2006 AVI Workshop on Beyond Time and Errors: Novel Evaluation Methods For information Visualization (Venice, Italy, May 23 - 23, 2006). BELIV '06. ACM Press, New York, NY, 1-5. DOI= http://doi.acm.org/10.1145/1168149.1168151


Evaluation References