Building the model
Click the Generate rosters menu. Until you have a project started, nothing on this menu is functional. Once you have a project, and have loaded data on individuals, the first two options (Report domains and Analyze) will be active.
Click Analyze now. Parity Builder will process your data (which may take a few seconds, depending on the size of your project and the speed of your machine) and will then write a cluster report in the Messages tab. (You may need to scroll the messages tab to see it.) The cluster report tells you what Parity Builder found when it examined your individuals and their attribute values. A cluster is a group of individuals who all have the same attribute values. In this case, that would be children of the same gender and grade level, either all experienced or all inexperienced, all with the same skill ranking, all from the same neighborhood. Individuals who are involved in one or more restrictions are automatically assigned to a cluster by themselves (called a singleton). Individuals who are not restricted but represent a unique combination of attributes will also wind up as singletons.
In addition to reporting on the number and size of clusters (which you can generally ignore, unless you are curious), Parity Builder also reports the attribute values it found among the individuals. Unlike the Attribute Domain Analysis, which merely lists the values of each attribute, the Cluster Analysis reports the distribution of attribute values. For each categorical and affinity attribute, it lists all the values it found and the percentage of individuals with each value. For each numerical attribute, it reports the range of values (minimum, maximum) it found, along with the number of distinct values and the population average (mean).
Having already cleaned up the data, our cluster report shows plausible results for all attributes. There are 59 clusters, the largest containing 10 comparable individuals, the median cluster containing just two comparable individuals. There are 26 singletons, which are the eight children involved in restrictions ("forced singletons") and another 18 apparently unique individuals.
The next step is to check solver settings, after which you will be ready to solve the model.
| previous: Enter restrictions | next: Adjust solver settings |
