Selecting Polygons based on multiple POI with radii using ArcGIS for Desktop?

Question

I have a polygon layer of suburbs (with population density) and I have separate layers of POI such as hospitals, government buildings etc. What I'd like to do is give each POI a radius based on how important I think they are (such as a private vs government hospital - from the attribute table). Then I'd like to highlight areas where the different POI radii overlap and lastly export the suburb the overlapping areas belong to as a list (it should include a percentage of the total suburb the overlap falls into, and the distance to the nearest hotel for instance).

Is there any way I can build a query that can achieve this in ArcMap 10.3? I need a fairly efficient way of doing this as there are several hundred POI (within each category) and about 7000 suburbs. I've tried using the select by location tool but what I described above seems to be more complex than it can handle without some manual adjustment possibly via python code.

Answer 1

You're trying to do far too much with a single operation or query. You need to break the problem down into smaller components, and possibly create a script or model to automatically run all the steps.

1. Buffer points, do not dissolve. This will give you a single buffer polygon for each point, some of which may overlap. It is possible to buffer based on an attribute field; this is covered at other questions here. You'd have to have that as an attribute though.

2. Union or intersect the buffers to themselves to isolate areas of overlap. This will create duplicate shapes, one for each that contributes to an area of overlap. At this point you'll need to decide if you want to continue with them separate or delete/merge the identical ones in a manner that lets you keep track of what contributed to them (ie, two different kinds of hospitals overlap). There are various was to do this, also covered here at other questions. Some involve the Dissolve tool.

3. Union the buffers to the suburbs. You'll get polygons that have attributes from all the input layers. From here you can determine the area percentages of the resulting polygons to their parents.

4. Further analysis may require other tools. For example you mention distance to nearest hotel, which would require a hotel point layer and use of a Spatial Join or the Near/Generate Near Table tools. This is also covered at other questions here.

It's up to you whether you want to operate on all the data at once, possibly combine your POIs to a single file with a category attribute, or iterate through them one at a time in a model. Efficiency is up to your process design and what that process needs to accomplish.