If I understood correctly then this was considered as a not necessary step.

Now given the Twitters nature then is there such a model that is even approximately accurate?

Or in other words, what relevant data Twitters collects and how accurate it is?