Three Dimensional Data Displays
A MultiGraf is essentially a three-dimensional sphere containing variables (called “attributes” that are shown as lines through the center) and data points (objects) within the sphere. The labeled end of an attribute line (arrowhead) is generally where the largest numerical value for that attribute is located. The angle between the labeled ends of any two attributes indicates the relative degree of correlation between those attributes. Attributes that are positively correlated will have an angle of less than 90-degrees between them. Attributes that have little or no correlation between them will be perpendicular to one another. Attributes with negative (inverse) correlation will have angles greater than 90-degrees. Thus ALL the relative correlations between all attributes can be seen at a glance.

Look at this MultiGraf for 5 cost-of-living attributes and 9 U.S. cities plus the U.S. average for all cities. Four of the five attributes are closely related because their labeled ends are at angles well less than 90-degrees. However, the Utilities attribute has essentially no correlation to Groceries and somewhat negative correlation to the other 3 attributes.

The position of a data point within the MultiGraf sphere indicates its comparative value relationship to all other data points for ALL of the attributes displayed. Imagine perpendicular lines drawn from all data points to an attribute. Those intersection points along the attribute provide a quick visual comparison of the data point values. In the example above, Sacramento has the highest comparative cost for Utilities while San Diego has the lowest, followed closely by Boise, Idaho. Thus, if a company were trying to determine where to locate a new technology plant where the cost of utilities was a major operational cost component, San Diego and Boise instantly appear to be better candidates (lowest comparative utility costs) than Sacramento (highest comparative utility costs).

More Information
For more detailed information on the MultiGraf multivariate analysis tool and its capabilities, including the Windows desktop version of this application, please contact NCOE. If you include a sample of your data in spreadsheet format (MS Excel preferred), an NCOE representative will get back to you with information about how this unique new data analysis tool may be able to help you better analyze and interpret your multivariate data.