I totally wasn't expecting to immediately duplicate this with a fidget spinner and the fluorescent light in my office. What a neat effect!
My technique: I hold the spinner quite close to my eye, so that it occupies nearly all of my field of view, and look through the spinning part at my ceiling light. If the spinner is very fast or very slow, I don't see any change in the hue of the light. However there's an intermediate rate where the back part of the fluorescent lamp takes on a patchy blue tint with colored edges. The tinted region is irregular and unpredictable in the same way as a pressure phosphene, if you're familiar with that better-documented effect, and has kind of the same "feel" to it.
Trying to photograph the effect just gives the rolling shutter problem, so that's no use. (No color change in the photo, either.)
At the rotational speed where the color change effect is the most pronounced, I estimate the frequency at which the spinner blocks nearly all of the light to be 10-20 Hz. (How? I let the thing hit a fingernail and listened to it tap; it sounds like thirty-second notes, four to a click, at a good presto tempo.) I've read elsewhere that the human eye can process images at 10-15 Hz. So some sort of biological stroboscopic effect, where the eye has just enough time to begin to respond to the reduced light before the light returns, seems plausible to me.