Logistic regression is definitely not what you need. It models the likelihood of an outcome of a dichotomous variable (yes/no) based on a bunch of predictor variables. So, for instance, the odds of a tweet being longer than 20 characters based on time of posting, language used etc. As already pointed out by PGStats, you need to specifiy what it is that you want to do with your data. Then we might be able to help out.
... View more
