soiltechproject / fsm-docs

https://www.soiltechproject.org/
3 stars 0 forks source link

[Discussion] Spreading cLHS generated sample points #3

Open Samuel22 opened 4 years ago

Samuel22 commented 4 years ago

Some feedback from FarmLab, the cLHS points are bunched. When sampling multiple paddocks, this can result in samples being contained within a single paddock. Suggest two 'options' for the user:

KipCrossing commented 4 years ago

Good point. The current python script runs all the columns by default and, via the CLI, we exclude the lat/lon columns. In the short term, for FarmLab, we could simply not exclude these columns.

Looking forward, however, this won't be so obvious from the scala interfaces perspective @correllink . This could be done like:

val samplesLocations = samsFarm.GenSampleLocations("clhc", 10, zones = [paddock01, paddock02], attributes = ["ndvi", "uGamma"], includeLatLon = true,  includeZones = true)

Remember that paddock01 and paddock02 would be zones with type="paddock"


In the back end, for the paddock zones, do we simply number them and use that column OR do we implement Objective Factor 2 that handles categorical data?