Transcript shneiderman
Dynamic Queries for Visual Information Seeking Ben Shneiderman Jin Tong Hyunmo Kang Cmsc838 Sep. 28, 1999 Outline • • • • • • • • • Dynamic Queries Examples of DQ Applications Advantages of DQ Disadvantages of DQ Enhance DQ via Movable Filters (Magic Lens) Video Clip of Magic Lens Boolean Queries by Composition Example of Query Composition Conclusion and Critique Favorite Sentence • “Visualization offers a method for seeing the unseen” Dynamic Queries • Interactive user control • Visual query parameters adjustment • Animated visual display of query results Why They Are Good • For novices: • For power users: - Don't have to learn SQL - Helpful in finding patterns - Avoid syntax errors - Explore and discover - Natural, aid comprehension Home Finder Home Finder (Text) UNIX - Ls Chemical Table Dynamaps Filmfinder Global Change Master Directory User Study Results Advantages • Visual presentation of query components • Visual presentation of results • Rapid, incremental and reversible actions • Selection by pointing (user interface improvement: what about voice command) • Immediate and continuous feedback (related: tight-coupling of DQ filters) Disadvantages and Research Directions • DBMS and display related performance problems * Data accessing algorithms * Display/screen management • User interface (domain dependent) Disadvantages and Research Directions (Cont.) • GUI issues (widgets, representations, etc) • Input methods • Novel user interface for complex queries Filter/flow Map Enhanced Dynamic Queries Via Movable Filters Ken Fishkin Maureen C. Stone Restrictions of Dynamic Queries (Motivation) • The number of attributes is limited by the number of selectors • The effect of combining slider filters is strictly conjunctive • The effects of the selectors are global • The number of selectors is fixed in advance Enhanced Dynamic Queries Via Movable Filters • Combining the two techniques : The starfield display, the movable filter • Enhancing the starfileld display by augmenting it with the flexibility and the functionality of the movable filter Boolean Queries By Composition • Lens L=(F, M) - F : filter Describing the output calculation for the filter on some datum - M : boolean operator Describing how that output is combined with the output from lower filters Example of Composition • L1=(F1, OR), L2=(F2, AND) - L1 over L2 (F1 OR F2) - L2 over L1 (F2 AND F1) • N=(NULL, NOT) : inverting lens • Compound lens - (F1 AND F2) OR (F3 AND F4) Examples : - Database : US Census Data - Lens Manager Server : X Window System Example of Composition Example of Composition Alternate Views Simultaneous Multiple Views Boolean Filter Extensions : Real-valued Filter Extensions : Real-valued Filter Missing Data Missing Data Conclusion • Expressive yet easy to understand • Powerful queries(boolean and real-valued) • Visual and semantic transformation of the data (callout, magnification, missing data, sorting, and so forth) • Wide range of interface operations (clickthrough tools) Critique • No statistics on the usability tests • Need rapid search & rapid graphical display • Application specific programming Favorite Sentence • “There is a tension in the database query systems between providing expressive power and ease of use”