I never realized the frequency changed based on load until I saw a video about everyone starting up their kettles at the same time causing them to have to fire up extra capacity. I guess it makes sense, since most power plants are based on generating AC from spinning turbines so any extra physical resistance is going to make it harder to turn, so it will turn slightly slower. All power plants will more or less sync due to magnetics. Each one would need to start pushing a little harder more or less in sequence to speed them up again.
That data logging is neat, would be kinda neat to set something like that up at home for fun.
Simpler than logging, if you want to see it in real time, get a 2 channel scope, get a crystal locked 50/60hz signal on 1 trace, and, probe the mains through a cheap ac transformer on the second channel. Watch you mains slowly scroll in one direction, then slow down, then scroll in the other direction, every few minutes. Well, it's like that for Hydro Quebec here in Montreal. The average over a few hours is corrected to be 60hz exactly, but, every few minutes, it drifts above and below 60hz visible on my scope with the above setup.
I believe in North America, recently, our 60hz grids, if the power is sold across province/state borders, our 60hz needs to be tuned to a GPS referenced clock, so it has been said on a PBS documentary about our power grid. I'm not sure how recent or accurate this info may be.