I want to cluster many points (around 1.000) into even sized groups (13 points each group). There is no attribute defining that groups, it should be grouped based on spatial distance.

I tried to work with "Clusterpy", "Concave Hull" and the long road via grids like in the following pictures. I made a four different grids (big to smaller ones) and the idea was to intersect the grid (starting with the big one) to intersect with the point layer. If there is the right number (13 points) or less but in the "surrounding area" not one point more as a single one - then that intersected point layer may be the first group. Then the progress of finding those spatial related groups would go on and the grid would get smaller and smaller. after the first grid i decided that this is not the ideal way to find a solution.

enter image description here

enter image description here

I also tried using "Shared Nearest Neigbhor Clustering" but with "13" or "8" as a number of Neighbors and only based on location every feature has the ID "1" in the resulting Output layer.

What would be the best way to split a point layer into even sized groups (same feature amount) based on spatial proximity?

I use QGIS 2.18

  • 1
  • I don't have an attribute like "population" as in your related question that would help to group the points. It should be based only on spatial proximity. – DirkB. Aug 25 '17 at 9:49
  • Make population = 1 e.g. gis.stackexchange.com/questions/250113/… – FelixIP Aug 25 '17 at 9:52
  • Those links aren't really helpful. First related question is mainly arcgis based and the described process is based on a field, second is also only 'related' but not similar. – DirkB. Sep 3 '17 at 13:29
  • Best of luck doing it your way – FelixIP Sep 3 '17 at 23:01

Your Answer

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

Browse other questions tagged or ask your own question.