Open peterdudfield opened 2 years ago
Would be interested in your thoughts @JackKelly and @jacobbieker
I would probably be inclined to leave it for more random systems. If you want only PV systems in a certain area, then you can chain together cropping a spatial area of the PV data, and then choosing N PV systems. It might be worth having a separate one that chooses the N closest systems, but I think that would just be almost the same.
Yea, i thought it would just be an option in that data pipe, either random or closest, like method='random' or closest
Ah yeah, in which case, yeah, adding it makes sense to me then, we probably want to get the N closest ones for like pseudolabelling as well
ok, ill give it a go
Detailed Description
When choosing PV sytesm for a example, option to only choose the closest ones
Context
if there are 1000 system in the system, and we want to only get 16 systems, maybe its best to get the PV systems that are closest, on the other hand, having a random selection means we could have lots of PV systems from a wide range
Possible Implementation
put something when we sleect the pv systems in here