I have a dataset consisting of geo points of two classes. For simplicity we can call one class "successes" and the second one "failures".

What I want to do is to find regions (areas, polygons) where ratio of failures to successes is relatively high.

If I had one class, I would do something like DBSCAN or OPTICS to find clusters.

I thought about separating whole region into smaller pre-defined sectors, computing ratio of failures to successes in each of them, and then to do LISA (Local Indicators of Spatial Association) clustering on them, but then my result will depend on the way I separate region into sectors.

How can I approach this problem?

Also, what if I wanted to divide whole region into sectors of several classes ("high probability of failures", "medium probability of failures" etc)?

I generally use Python and PostGIS but open to other tools. I dont have ArcGIS license.

  • What have you tried? If trying to do this with python, please include a snippet of the code you have tried and outline what happens when you try that code
    – Midavalo
    Sep 6 '16 at 18:25

Convert both sets of data to a raster layer using Point Density. For example, successes per acre. Now you will have a point density raster for failures and a point density raster for successes. Divide successes by failures. Reclassify into the groups you desire and convert to polygons as desired.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.