Scatterplot of Anomaly Index by Variable Impact
The tables contain a lot of useful information, but it can be difficult to grasp the relationships. Using the saved variables, you can construct a graph that makes this process easier.
- To produce this scatterplot, from the menus choose:
Figure 1. Chart Builder dialog box - Select the Scatter/Dot gallery and drag the Grouped Scatter icon onto the canvas.
- Select Anomaly Index as the y variable and Reason Variable Impact Measure 1 as the x variable.
- Select Peer Group ID as the variable to set colors by.
- Click OK.
These selections produce the scatterplot.

Inspection of the graph leads to several observations:
- The case in the upper right corner belongs to peer group 3 and is both the most anomalous case and the case with the largest contribution made by a single variable.
- Moving down along the y axis, we see that there are three cases belonging to peer group 3, with anomaly index values just above 2.00. These cases should be investigated more closely as anomalous.
- Moving along the x axis, we see that there are four cases belonging to peer group 1, with variable impact measures approximately in the range of 0.23 to 0.33. These cases should be investigated more thoroughly because these values separate the cases from the main body of points in the plot.
- Peer group 2 seems fairly homogenous in the sense that its anomaly index and variable impact values do not vary widely from their central tendencies.