Page One

Campus Computing News

Simeon Support to be Phased Out

The Force is just around the corner ... really!

Faculty/Staff NDS Directory Update

Today's Cartoon

RSS Matters

SAS Corner

The Network Connection

List of the Month


Short Courses

IRC News

Staff Activities

Subscribe to Benchmarks Online

Research and Statistical Support

SAS Corner

By Dr.Karl Ho, Research and Statistical Support Services Manager


In this short article, I will introduce a tool that is rarely in the spotlight, but is a useful one for exploratory data analysis and data visualization. The Spectraview module is available in both the Windows and UNIX/ XWindows versions (version 8). The multiple window display interface of Spectraview enables users to visually explore trends and relationships represented by the data. For example, environmental scientists can use this tool to investigate and identify the level of sulfate concentrations at different longitudes and latitudes. Controlling for other variables, scientists can visualize and animate the changes of the dependent variable as caused by the independent variable. The following example illustrates a study of sulfate concentration as affected by layer above ground, latitude coordinate and longitude coordinate.

To activate the Spectraview module, select Solutions --> Analysis --> 3D Visual Analysis. 

Before analysis can be conducted, data have to be loaded into Spectraview. It takes only SAS data sets.  Click on the data button and select the data set from the SAS libraries:

Spectraview example

Once a data set is chosen, variables in the data set will be displayed for selection.  Click on the Read Data button after selecting the variables of interest.

Spectraview example

By default, Spectraview presents a four-window display, to accommodate simultaneously four variable interactions. The main response variable will be displayed at the upper right window. The rest of the windows present the relationship between any two independent variables. 

Spectraview provides a spectrum of visualization techniques including: 

  1. cutting planes - produces slices of spatial data

  2. Direct volume rendering - creates a two-dimensional image of the entire volume of data points with transparency.

  3. Isosurface - produces a three-dimensional surface by connecting all the data points with one response value; and

  4. Point cloud - displays response values with colored markers, showing individual data points.

The following 3-D chart is created using the isosurface technique to plot the relationship between the three independent variables by identifying a range of values of the response variable (sulfate concentration):

Spectraview example

A full-fledged analysis can be conducted by plotting the relationship of all variables and their individual impact on the response variable.  

Spectraview example

The researcher can choose different levels of each of the independent variable to detect the highest and lowest level of sulfate concentrations using the four-window interface.   

Another advantage of this module is the availability of the preset data filtering options, that allows for smoothing or contrasting data. The following example illustrates another data exposition that animates the changes of data by changing one independent variable. Click on the chart to start the animation (or click here to download the AVI file):

Spectraview example