Skip to main content

image2image1Hi All,

 

I'd like to split our entire road network into 5 groups.

  • Each group would cover approximately the same total length i.e. (Total length of all roads)/5
  • The roads in each group would be spatially clustered/grouped together
  • Ideally i would not be snipping any road segments

 

I've tried to accomplish this by using the PointClusterer and feeding in the roads center points or 10m interval points but when merging the cluster points back to roads, the grouped distances are too varying.

 

Is there a better approach to this?

 

Thanks



It kinda reminds me of the Garbage collection excercise (runs through that entire chapter) from the FME Desktop Basic training, which used the SpatialSorter custom transformer.

 

I think you're on the right track by analysing points rather than lines and my suggestion would be to see if you could use a process similar to that exercise.


Reply