Being a QS geek myself and partner in a QS based startup, I am always interested in new products. However, it seems to me that 90% of people doing anything in

Being a QS geek myself and partner in a QS based startup, I am always interested in new products. However, it seems to me that 90% of people doing anything in this space have the same idea: track as much as possible and then try and see if there's anything interesting there. "44% more likely to check in when you climb more floors" -- big whoop! I suspect that this is exactly the level of insights that you are likely to gain from that sort of approach. You will discover that you are indeed less productive if you don't sleep well or that you tend to have headaches after drinking large amounts of alcohol. Exploratory data analysis is appealing because we like the idea of discovering something completely unexpected and revolutionary but I think confirmatory analysis (the boring one where you form a hypothesis and set up an experiment to test it) is much more likely to be useful.