I've been wondering for awhile why we can't use this $30
PublicLabs DIY spectrometer. You connect it to their
SpetralWorkbench. Calibrate it to a CFL (some consistent mercury vapor signature present in all CFL bulbs, I believe is their theory.). We could then "calibrate" our graphs relative to each other by taking a graph of noon-time sunlight. It would be a process.
The workbench lets you export the data to CSV for a spreadsheet which can be graphed/smoothed. We could all use the same spreadsheet which would do the relative adjustment using the two exported datum. (A spectral reading of an LED light, and a reading of noon-time sunlight taken with a spectrometer that hasn't changed or been CFL calibrated since that sunlight measurement.).
It would be some work. I'm not volunteering to do it. Just throwing out an idea. I have this spectrometer and would be willing to participate. (I can definitely document some gotchas so we'd all use it the same way.).