Chapter 3 Visualization
Section 3 Visualization Techniques
Page 4 Temporal and Mulitidimensional Attribute Techniques

Objectives

The objectives of this section are:
to introduce the basic concepts in data visualizationto explain the various visualization techniques
to understand in which situation a particular technique is used
to introduce higher dimensional visualization techniques that exist

Outcomes

By the time you have completed this section you will be able to:
choose a visualization technique based on the dataset attribute
create a scatter plot, histogram and stem and leaf plot
list common visualization techniques

Visualization of data with spatial/Temporal attributes

We sometimes have data that has spatial or temporal attributes by this we mean that a given data set can consist of a set of observation on a spatial grid and can vary at different points of time. Visualizing these attributes cannot be done with the techniques previously discussed and so other techniques must be employed.

Contour Plots

Visualization of a 3D data is a contour plot. As mentioned above there are situations in which we would like so view 3D data by this we mean that two attributes indicate a position in a plane and the third attribute has a continuous value (for example time). In this particular case it would serve us well to use a contour Plot to visualize the data. A contour plot divides the plane into separate regions where values of the third attribute are more or less the same.

Surface Plots

Just like the contour plots, surface plots use two attributes for the x and y coordinates and the third coordinate is used to indicate the height above the plane. Such graphs necessitate the value of third attribute to be defined for all combination of values for the first two attributes. Problems arise when the surface is too irregular, to circumvent this issue the plot could be viewed interactively.


Other techniques used for visualization of spatio-temporal data include vector filed plots and lower-dimensional slices.


Visualization of data with many attributes

For a data set that has many attributes, techniques exist to visualize these many dimensions but all of these techniques have limitations of their own that need to be understood so that the proper technique is used to provide us with the aspect of the data that we care the most about.  Some of these techniques include Matrices, Star Coordinates, Chernoff Faces and Parallel Coordinates.

Star Coordinates and Chernoff Faces

These techniques encode objects as icons or glyphs. Imagine for a second you wanted to decode an Egyptian hieroglyphic, one would require a key in order to figure out what each symbol represents and star coordinates and Chernoff faces are visualization techniques that build upon this idea. The main idea of star coordinates is to map the data objects onto a polygon using the axes as the attributes. Chernoff face is as the name suggests and each attribute of the data object is mapped to a feature on the face.