I find it also very interesting and worth to "investigate". However, one should not get lost in that "chaotic" behavior; such behavior is actually expected for such a problem. When you illustrate prime number distributions you also will not find any clear pattern. Despite the fact that there is no clear pattern deducible for prime numbers, proofs related to/about prime numbers exist. It is well possible that one can solve the Collatz Conjecture but never explain this "chaotic" behavior that you have mentioned.