Open simonrolph opened 2 years ago
I'm sure there are spatial clustering algorithms that could be used
This looks like a fairly straightforward clustering approach: https://gis.stackexchange.com/questions/17638/clustering-spatial-data-in-r
Getting the threshold is going to be a bit of visualising and seeing what seems sensible.
But a question is do sites need to be consistent across observers? eg. can we apply clustering on the observations from each of the observers? Or do we need to apply the clustering on the whole set?
This would be good to feed into the core package
Tried this one on all records (not just one observer) and it tried to make a 300gb vector... need a more efficient way of doing this
This looks like a fairly straightforward clustering approach: https://gis.stackexchange.com/questions/17638/clustering-spatial-data-in-r
Getting the threshold is going to be a bit of visualising and seeing what seems sensible.
But a question is do sites need to be consistent across observers? eg. can we apply clustering on the observations from each of the observers? Or do we need to apply the clustering on the whole set?
This sounds like a good idea. Maybe we can chat in more detail about this
In each case for running in a different country we'd need to find their grid system - what about if we wanted to run across national boundaries etc.
Is there a better way of grouping records by location which can be more easily applied in other settings?
What is a site?