It does not address the problem of plotting large numbers of sites.
I don't have a solution, ggplot doesn't easily handle anything that can't fit into a long form data format. Converting to something like a NetCDF file, then using packages like tmap to plot, might work, but what value does a plot of a similarity matrix have for millions of sites?
The way forward might be to just randomly subsample sites for plotting. The general clustering trendwill be visible.
3 provides a path forward to exceed 45 000 sites.
It does not address the problem of plotting large numbers of sites.
I don't have a solution, ggplot doesn't easily handle anything that can't fit into a long form data format. Converting to something like a NetCDF file, then using packages like tmap to plot, might work, but what value does a plot of a similarity matrix have for millions of sites?
The way forward might be to just randomly subsample sites for plotting. The general clustering trendwill be visible.