Mapping-Tools logo MPCluster

Cluster Analysis for Caliper Maptitude®

Guided Tour of MPCluster: Adjusting the Display Settings

The MapPoint version of MPCluster was restricted to displaying the clusters as simple map annotation. Maptitude supports a much richer set of data capabilities, and these are reflected in MPCluster's display options.

Example of clusters with boundary and centroid layers

Create new layers from the clusters

The Write clusters to layers option is the closest output option to the MapPoint version's "Annotation" output. This option creates (or replaces) two new layers to display the new clusters. One of these layers is an area layer and draws boundaries for each and every cluster. The second layer is a point layer that marks each cluster's center with a solid triangle. Both layers use matching colors for each cluster.

The image on the right shows cluster layers displayed in this manner but with the input data hidden for clarity.

The cluster boundaries are drawn as convex hulls around each cluster's data points. Note that Hierarchical clusters might be slightly concave, resulting in cluster boundaries that appear to overlap.

Input data colored as per their cluster allocations (click for larger view)

Using cluster allocations to color the input data

It is possible for MPCluster to color your input data according to the clusters that it has been allocated to. This is one way to add clarity when the cluster boundaries appear to overlap.

This option is controlled with the Write allocations to a data view option. This creates (or replaces) a data view that lists all input data points along with their cluster allocations. Further options let MPCluster join this data view to the input view, and to apply a theme. This theme colors each data point according to the allocated cluster. These colors match the colors used by the boundary and centroid layers. The image to the right shows an example without the boundary layers.

The Professional license adds the ability for MPCluster calculate overlay values for each of these clusters, writing the results to the new cluster layers. These overlays are calculated by summing or counting fields in other layers. For example, you might count the number of customers or sum the total sales.

Next, we look at the MPCluster's Excel output.

Buy MPCluster Now!