To avoid Paralysis Analysis, what helps me move along is do it dirty first, make it work, learn as much as you can at this stage, then commit to an idea.
I find that when I make it work, it becomes easier to organise a plan to tackle the actual implementation, and most of the times the “dirty” implementation is more than enough for solving the current task at hand.