Hi everyone. I've posted screenshots of my graphs and the CSV for the last 5 months or so; I spent much of it alternating days using 1.5mg of melatonin. The results seem to be that it decreased my sleep need by ~1 hour, didn't affect length of deep sleep, and reduced length of REM sleep. See http://www.gwern.net/Zeo#melat.....n-analysis
If anyone wants to do a better analysis with the CSV or share their own data series, that'd be interesting.