Currently individual datasets have individual constructor parameters, for example:
# Temperature data options
frank_temps_file::String = "Data/FranksTemps.mat"
bisicles_temps_file::String = "Data/antarctica-bisicles-xyzT-8km.nc"
Even when they are the same "type" of data (Temperature, in this case).
Works for now, but ideally this is simpler, and a "type" of input data is controlled by a single parameter. This could get out of hand when there are many different possible data sources for a data category.