VTDP Dynamic Queries

 

In some data set all entities have the same set of attributes.  The attributes define the data dimensions and each entity can be thought of as a point in a multidimensional space. A set of sliders is provided that can narrow the range on each of many attributes. Each slider adjustment is an epistemic action, narrowing the range of what is displayed. 

Example 1: A database of houses is shown on a map with a symbol for each house. The user adjusts sliders to select the price, number of bedrooms, bathrooms, square footage of living space, and distance from the nearest school [1].

Example 2: A scatterplot shows the set of stocks traded on the New York Stock Exchange. The axes represent market captitalization and annual increase/decrease in value. The user adjusts sliders to select price/earnings, monthly gain. Radio buttons are used to select industry sector.

The method has most often been used with scatter plots and to a lesser extent time series plots, but it can work with ranges displayed on maps, and with node link diagrams where restrictions can be placed on the nodes or links.

Visual thinking process with dynmamic queries

Display environment: A display with symbols representing entities drawn from a set of multidimensional discrete data, with a set of controls that restricts the range displayed on each of the data dimensions.

  1. User constructs task relevant visual query that can be addressed by viewing a subset of multi-dimensional discrete data defined by a hyper-box.
  2. Execute two visual queries on display (a)Is the number of targets small enough to make drill down feasible. (b) Is a pattern found or high relevance symbol visible?
  3. If as a result of 2b a high relevance symbol is found, execute an epistemic action to drill down for additional information. Usually the results will be presented in a different display window.
  4. Otherwise, execute an epistemic action, dragging a slider which causes the computer to adjust a range on a data dimension and display the modified subset of the data.
  5. Repeat from 2 until either task is successfully completed or abandoned.

Spotfire

The image above shows Spotfire a commercial tool that makes extensive use of Dynamic Queries [1].

Implementation Guideline: Ideally, update following slider manipulation is very rapid (<100 msec).

Data Guideline: If we assume that each dynamic query slider can be used to reduce the range selected to 10% of the original, then the number of objects that be interactively queried is > 10 to the power d where d is the number of dimensions. So five sliders will allow for 100,000 objects to be interactively viewed. If the goal is to get to a small number of objects for drill down then the final on-screen display should contain fewer than 10 objects.  

Visual Query Guideline: If additional attributes can be encoded in the data glyphs this can facilitate visual search for the sought after information.

References

1. Ahlberg, C., & Shneiderman, B. (1994, April). Visual information seeking: tight coupling of dynamic query filters with starfield displays. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 313-317). ACM.

 

 

 

    Visual Thinking Design Patterns are partially funded by the DARPA XDAPA project