Generate best fit polygon based on aggregated nearest feature attribute

1929
6
Jump to solution
03-05-2020 12:09 PM
AndrewDeLuca
New Contributor III

I am trying to figure out how to generate a boundary around features in a polygon feature class that is based on a count of a field in a point feature class tracking estimated population. The point feature class has a field for estimated population. I would like the tool (or model/script if I need to write one) to find the nearest x features from the point feature class to the polygon feature class that equal a set value based on the estimated population field in the point feature class. So if I input 200 it would iterate through the nearest point features until it reaches 200 for the attribute I have selected, then generate a minimum bounding geometry around those points. I am trying to create a service area of sorts, based on a needs assessment that is population based.

Any idea if something like this already exists? If not, ideas on the tools to string together to do it?

0 Kudos
1 Solution

Accepted Solutions
DanPatterson_Retired
MVP Emeritus

There is nothing out of the box that I know of, it will require scripting

View solution in original post

0 Kudos
6 Replies
AndrewDeLuca
New Contributor III

Hi Dan, I am not looking to make bins or boundary areas and then summarize the population estimate value. The workflow would be select the nearest points until you reach set value by summing the values in the est population field, then generate a polygon around the selected points. I do not believe those tools would accomplish that unless I am mistaken as each requires a set polygon. The polygons in question for my analysis do not encompass the points I am using for population estimates, rather all the points lie outside of the polygons. Sorry for the confusion.

0 Kudos
DanPatterson_Retired
MVP Emeritus

Ahhh, The question would then be how you go about selecting the nearest points. 

  • Do you have seed locations around which you want the clusters to begin?  That is, location is initially the primary concern, then aggregation proceeds.
  • Alternately, would a seed location be an existing location with a high value for your base parameter?  Location is secondary and distance to subsequent location is less of a concern
  • Would distance be the primary clustering parameter, followed by your population field values?

The pattern that you end up with will depend on your existing distribution of points and how you begin your clustering process. 

The first option in the list I provided, is solely a distance based clustering requiring no aggregation. 

The remaining options provide an aggregation framework into which you can aggregate attributes. 

Just some thoughts.

0 Kudos
AndrewDeLuca
New Contributor III

The first option is correct. It would go like 

  1. Input seed location
  2. Find nearest point & record population variable
  3. Is pop variable > input threshold?
  4. if not, find next point, add to population variable
  5. Repeat until threshold is reached
  6. Create minimum bounding polygon around all points selected during this process
0 Kudos
DanPatterson_Retired
MVP Emeritus

There is nothing out of the box that I know of, it will require scripting

0 Kudos
AndrewDeLuca
New Contributor III

I figured as much, thanks for confirming Dan!

0 Kudos