Example of Multidimensional Scaling

In this example, use MDS to construct a map of 28 cities in two dimensions that is based on pairwise flight distances. The data table that contains the flight distances is a distance matrix.

1. Select Help > Sample Data Folder and open Flight Distances.jmp.

2. Select Analyze > Multivariate Methods > Multidimensional Scaling.

3. Select Birmingham through Washington DC and click Y, Columns.

4. Select Distance Matrix from the Data Format menu.

Figure 10.2 Completed Multidimensional Scaling Launch Window 

5. Click OK.

In the Multidimensional Scaling Plot, hover over data points to view the row number or row label. The next 7 steps are to label and rotate your MDS Plot.

6. Select the Flight Distances data table.

7. Right-click the column Cities and select Label/Unlabel.

8. Select Rows > Row Selection> Select all Rows.

9. Select Rows > Label/Unlabel.

10. Select the Multidimensional Scaling Plot.

11. Click the Flip Horizontal button.

The Flip Vertical and Flip Horizontal buttons enable you to change the orientation of the MDS Plot. The MDS results are invariant to orientation. When the results have a known orientation, such as physical locations on a map, then you might want to rotate or flip your plot.

Figure 10.3 Multidimensional Scaling Plot