The part number on the graph is a LiPoly battery if you google it
There are problems trying to use voltage to measure state of charge (quite obvious from those curves in fact), that the voltage depends on more than just the state of charge - it depends on batt temperature, the current flowing (due to internal resistance), whether the battery is charging or discharging. It will also change as the battery ages (and internal resistance itself changes with age).
Some chemistries have very flat curves (lead acid is fairly flat, NiCd is very flat), and this makes it very inaccurate to try and judge S.o.C. except at extremes. NiMH and lithium chemistries do have reasonable slopes IIRC (and probably for the same reason, come to think of it).
If you can record actual charge flow history as well as voltage this ought to be much more robust at keeping track of S.o.C. and current capacity.