SS Gem is a bright example of the RV Tauri class, and more specifically of the RVa subclass whose mean brightness remains relatively constant. (The RVb class, of which a star like U Mon is an example, are characterized by long-period changes in brightness.) They're quasiperiodic pulsators with periods of several weeks to a few months, and are highly evolved stars, not far from the ends of their lives. SS Gem has been well followed by amateur observers for decades, although visual observations are dropping off as is happening with many other stars. It's also an example of "noisy" data that isn't necessarily. At first glance, this light curve looks like it has a lot of scatter, and it does, until you start looking at data from individual observers, visual and CCD alike. Careful examination of visual data will often show that over the long term, individual visual observers have a good handle on how the star varies, and the "scatter" in light curves is often due to offsets between observers, or differences in sequence. (And you can also see that CCD and DSLR observers, shown in green, have similar issues.) Most observers are careful with what they do, and researchers who use their data can get excellent results when using equal care.
One of our current projects at AAVSO Headquarters is to use statistics to measure how different observers perceive the same star. Early results show that the application of simple offsets to individual observers can reduce scatter substantially. Our next step is to see whether these offsets can be applied across all stars that an individual observes, and whether these offsets change with age or with the color of the star observed.