The iconographic technique can be used to provide a single and unifying view of a data set. This overall
view enables users to discover global structures in that data space. However, in order to allow a deeper
exploration of the data set, tools for navigating that data space must be provided. One way to support
navigation is by using other visualizations tools to produce other views which would be linked to the global
visualization provided by the iconographic technique.
Figure 3: Iconic technique integrated with other visualizations techniques
This course of research will examine the integration of other visualization tools into the iconographic technique as
a mechanism to facilitate the exploration of complex data sets. Two approaches will be examined. The first
approach will be based on the availability of certain other visualization techniques which users will be able
to select from a menu. These other visualizations will be used to display either the whole visualization
represented in the iconographic display or selected regions of the display. The visualization will be presented in a
separate window linked to the original iconographic display. Other multidimensional techniques such as Parallel
Coordinates [21], RadViz [20], and Scatter plots are being considered at this point, but
yet other techniques may also be studied and added in the course of this research.
Depending on the technique being integrated some interactions involving the linked representations will be
developed. The idea is to carry out graphical operations across the linked representations to investigate for
example, the effects of different mappings or how patterns in the iconographic representation are related to
patterns in other representations. That might require extending the techniques being integrated to support
the functionality required for such linked interactions. Figure 3 shows an iconographic display in which a
selected region of the iconographic display is being visualized using Parallel Coordinates and RadViz.
The
Parallel Coordinate representation shown in Figure 3 is an extended version of this technique in which sliders have been added to
each axis. These sliders can be directly manipulated for selecting or browsing through the data being
represented. Another functionality included in this extended version of parallel coordinates is the possibility
of swapping axis by direct manipulation (dragging). Swapping axis in the parallel coordinate representation
might be equivalent to change the corresponding mapping in the iconographic representation, allowing for
dynamically exploration of different mappings.
Figure 4: A dynamical approach for integrating the iconographic technique with other visualization techniques. In this figure RadViz has been associated with the moveable box.
Another approach that will be explored is based on a moveable 2D box that can be moved and sized to
enclose a sub-region of the iconographic display. A visualization technique will be associated with the
moveable box. All the data inside the region defined by the moveable box will be visualized in a separate
small window using the visualization technique associated with the box (see Figure 4). By moving the
moveable box over the iconographic display the whole iconographic representation can be dynamic
explored providing the basis to dynamically investigate how both representations (iconographic and the
visualization technique associated with the moveable box) relate to each other. This approach could also
be extended in such a way that analysis tools or filter operations can be associated with the moveable
box.