I am working on a project with a huge amount of data. As always, my task is to find the most interesting parts of the data and the appropriate model to use. The issue is that the path from here to completion is non-linear. There are many possible paths and I'll need to try out many of them. I'm sure this is familiar to you academics out there (and to others as well).

The problem is managing time while I'm doing this. What happens in practice 1.I get analysis paralysis and can't do anything because I can't decide on which route to try first.
2. Once I do try something, I stay on it too long. I spend too much time learning software commands, etc. for the benefit I get. And then I neglect other routes.

Has anyone found a way around this?

Thanks!