Organizer:

Steve Eick, eick@research.bell-labs.com

Speakers

3:45 p.m.
User Interfaces of Interactive Statistical Graphics Software
Martin Theus, VIAG Interkom
Keywords: Graphical User Interface, Interactive Statistical Graphics, Interface Design

Early graphical user interfaces were confronted with the attitude "Real men don't use mice". In many cases this skepticism was easy to understand, since the design of early GUI (graphical user interfaces) based systems was even more cryptic than their text based counterparts.

Today, Apple and Microsoft have widely familiar interface design standards. Which are similar but still different enough to be confusing. X Window System users are accustomed to being confused, since each application is likely to have its own idiosyncratic design style.

This talk will give an overview of some do's and don'ts in user interface design. Various examples will show that the implementer's intuition is often not enough and show the need of system atic approaches to efficient user interface design techniques.

Statistical graphics software benefits strongly from these design principles. Various interactive statistical graphics packages will be investigated to show how closely we can follow these principles in order to achieve a "good enough" design.

4:15 p.m.
GUIs for the Geographer: Some Experiences with Interactive Geo-graphics
Jason Dykes, University of Leicester, U.K.
Keywords: cartography visualization spatial statistics virtual environments panorama

Techniques developed in cartography convert the multidimensional and detailed properties of spatial phenomena into limited representations for a particular purpose. Advances in computing technology have lessened these limitations. Computer maps can change with rapidity to suit the use and user allowing the map to be reconsidered as an interactive interface to spatial information that provides successive user controlled views of a data set. The results include impressive and realistic graphics that appear to show information in three dimensions and interactive tools that enable users to query, slice, view and review the properties of spatial phenomena.

Such representations are particularly suitable for 'visualization' the iterative process of visual exploratory analysis. Whilst maps that are suitable for visualization can take advantage of some of the principles of traditional cartography effective 'maps' may be very different to those produced in the static media.

Key visualization interface issues involve integrating information, 'map' flexibility, filtering information and using appropriate affordances to guide users towards useful and insightful interpretations. Examples will be demonstrated that grapple with these issues in the context of particular applications:

'cdv' demonstrates the visualization of enumerated data and displays spatial statistics in novel ways.

'panoraMap' provides a virtual environment of linked geo-referenced panoramic images in combination with a map-based interface to statistical and qualitative spatial data.

Experiences gained from developing and using these interfaces will be used to explain why they, no doubt blatantly, omit the dos and embrace don'ts prescribed in the previous presentation!

(Please follow this link for more details, software etc.)

4:45 p.m.
Information Visualization: A Taxonomy of User Tasks
Stacie Hibino, Lucent Technologies
Keywords: data analysis process, information visualization, task analysis

Previously, most of the work in information visualization has focused on providing unique views or novel frameworks to aid users in exploring or accessing data. Very little work, however, has been done to support users through the full analysis process--from the raw data to the final results. But what tasks do users perform when analyzing data using an information visualization (infoVis) environment? A task analysis of experts' use of an existing infoVis environment was conducted to examine this question. Results indicate that users spend a significant amount of time working on tasks outside of data exploration--tasks such as preparing data, collecting results, and gathering evidence for a presentation. This study provides a working taxonomy of key data analysis tasks that (expert) users perform when using an infoVis environment and highlights the tasks that users indicate are the most important for their overall analysis goals.

5:15 p.m.
Antony Unwin (discussant), University of Augsburg