Guided Tour of MPCluster: Overview
MPCluster finds natural groups (i.e. clusters) in Caliper® Maptitude® data layers. These layers are typically point layers, but MPCluster can also work with the area layer centroids. The example map on the right shows the locations of restaurants in the UK. In this data, a cluster would represent a large number of restaurants in a small space. Therefore a restaurant supply company looking for new sales locations or new sales territories, would be very interested in finding clusters in this data. This map is included in the example archive.
MPCluster's Main Panel
This is MPCluster's main panel. It lets you choose the dataset to use for the input data, set the clustering algorithm and options, set the display options, and to start the processing.
The clustering options let you define the minimum and maximum cluster size, as well as the maximum number of clusters to find. These options are explained further on the next page.
Calculated clusters can be drawn as new boundary and center layers. Input data can also be colored according to cluster allocations using a joined data view. The display options are used to set these options.
After setting all the options, press Start to start the clustering process. Processing status is displayed using the progress indicator.
And here are the results. Each cluster is marked with a central triangle, and a red outline has been drawn around all of the cluster's component data points. All component data points are colored to match the cluster's color.
MPCluster has successfully identified the larger cities, as well as a number of smaller town/rural clusters which also meet the specified parameters.
Next, we look at how to set limits on the cluster sizes.