Hot Spot Analysis with Partitioned or Data with Breaks

[LarryGaudieri]

Hi,

I have a large data-set consisting of ±70K incidents. The scale is county-wide minus municipal limits (which creates breaks or holes through the data).

My question is: has anyone had this experience using the tool? If so, what is the approach? Should I partition or segment the analysis or is the tool intelligent enough to account for large areas without data?

Thanks for any help,

[XanderBakker · Accepted solution]

Hi Larry,

The (optimized) Hot spot analysis does not account for large areas without data. It is not an interpolation, but a statistical analysis of the number of incidents. If there are no incidents, clustering will be low, which will not result in a hot spot.

Some reading:
How Optimized Hot Spot Analysis Works
How Hot Spot Analysis (Getis-Ord Gi*) works

Kind regards,

Xander

[LarryGaudieri]

Thanks Xander,

Should have mentioned that I was still using 10.1

I did some looking and experimented with the interpolation tools and they accommodate breaks and voids but I'm still trying to understand the pattern of the result. I'll keep poking it...

Thanks again