Would a hot spot analysis work on a scattered island group?

[TaraLal]

I have point data (170 points) for the Solomon Islands and I'm trying to run a hot spot analysis. I'm facing a few problems-

1. the point data is partly on land within administration polygons and some fall off the polygons making it difficult to aggregate the data.

2. I can't find any hot spot analysis that's been run on islands that are this scattered - would this even work? 

Any thoughts?

Thanks!

[DanPatterson_Retired]

An image would help.

But one observation would be, if the observations are on separate islands (for instance), then they cannot be treated as a whole, but as separate entities