4

I've done the following process:

  1. take a set of points, each one representing a hotel with a field that says the "neighborhood" that the hotel is in (field is called "full_hierarchy")
  2. Performed SNN clustering on these points
  3. Created Voronoi Polygons using the clusters
  4. Dissolved the Voronoi polygons on the field "full hierarchy")

And I get this:

enter image description here

My problem is that there are bits of each polygon inside of other polygons, as you can see in my image where I have highlighted one of my neighborhoods.

Any ideas how I can get rid of these bits so that each of my polygons is complete with no "islands" in them?

I am going to be doing this on a mass scale, so I need a process that I can do on QGIS

Would also be interested in smoothing the polygons, if possible.

EDIT: Using the options recommended to be below when using the eliminate sliver polygons:

enter image description here

But the output layer looks exactly the same - nothing changes?

2
  • @Kazuhito Any ideas :)
    – Steven
    Commented Mar 2, 2017 at 21:07
  • No :) But potentially Vector | Geoprocessing tools | Eliminate sliver polygons or GRASS v.clean or something like Buffer + Debuffer ways. I myself would like to have answers from more experienced people.
    – Kazuhito
    Commented Mar 2, 2017 at 21:34

1 Answer 1

2

As @Kazuhito suggested the 'Eliminate Sliver Polygons' is the tool to use in this scenario. The tricky part is determining your cutoff point for feature removal. You will want to calculate the area for all the features in the layer. Using the area calculated find the value at which you would classify a feature as an "island"(sliver). In the 'Eliminate Sliver Polygon' select your layer and put in the area field as the selection attribute. Then set the comparison the less than or equal to the value that you have decided to be corresponding to the largest "island". The merge option will be based on your preference for what will match your study requirements. I generally merge with the feature that has the largest common boundary as it tend to maintain the integrity of my analysis better than the other methods. After that you need only designate a filename and location and you are good to go.

4
  • thanks for your help...I've added another image showing what choices I'm making when I use the eliminate sliver tool. With the choices I have selected, nothing to seems to change in the new layer created? Thanks for your help
    – Steven
    Commented Mar 6, 2017 at 20:37
  • @Kazuhito any ideas Kazuhito?
    – Steven
    Commented Mar 7, 2017 at 18:18
  • or any ideas why this isn't working from you Kingfisher?
    – Steven
    Commented Mar 7, 2017 at 18:58
  • I see you are working with a dissolved layer. It could be that the smaller polygons are being associated with the larger polygons. Try a single part to multi-part on the data and try the eliminate sliver polygons again.
    – Kingfisher
    Commented Mar 13, 2017 at 18:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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