# Create certain number of Clusters, equal sum of Z value

Divide polygons into *n* number of groups of equal counts with ArcGIS 10.2 is very similar to my question. But instead of equal number of polygons, I have multiple points (schools) with a z value (students in a special program). I am in 10.2. I have spatial analyst.

I need a certain number of groups (based number of program aides) that each have about the same sum of all their point's z value and are grouped spatially. The idea is that each aid goes administers the program at schools in an area and all of the aids have about the same number of students(z value) to help.

I have tried grouping analysis to see of it could get me close and no dice.

I know I can do this manually but I have a feeling this analysis will be required again for a different program so I would like to automate it as much as possible.

Attached is a proportional symbol map showing the range of z value and the general location of schools. I want to group it into 3 groups, same sum of z values.

EDIT/UPDATE

I think I will just visually (and with input from program director) pick three Central locations and create service areas using network analyst. Once I have finalized the appropriate service areas I will export them into a polygon shapefile and sum the number of z values for the school that are in the nearest service areas. Any schools left in a middle zone will be distributed to the group with the lowest z value sum.

• My solution to the post you are referring to designed to solve the problem you are talking about. Field [p2013] is you Z. Equal count per group is very least you can do Feb 8, 2015 at 21:49
• You are misinterpreting the intent and correct application of the Geits-Ord statistic ("Grouping Analysis"). The statistic is intent is to evaluate spatial clustering using an assumption of complete spatial randomness as a null. It is not intended to derive specified "n" clusters not is the Z value intended to specify this type of discrete process. Please read the literature on point pattern analysis and the Geits-Ord statistic specifically, and do not rely on the ESRI help document. Apr 4, 2016 at 15:33