One of my RC rabbit hole projects was to make a reverse probabilistic lookup from a users’s timezone (which is presumably set to local time) to locations. Nepa

One of my RC rabbit hole projects was to make a reverse probabilistic lookup from a users’s timezone (which is presumably set to local time) to locations. Nepal for example uses a +545 offset, and a one in 6 chance that you’re in the Kathmandu valley (based on population stats, but you could do better with internet stats). This works quite decently for small unique timezones, but gets fuzzier for the larger ones (China and India are just one timezone each). I then thought of using other Signals to augment the probabilities (such as the locale which does have its own issues). At that point it felt like a fingerprinting project instead of anything novel and I lost interest.