There has been a good amount of discussion recently about how certain people receive a lot of bullying, trolling messages on Twitter. Twitter has pretty basic tools for blocking trolls, but what if there was a tool that was more like a spam email classifier? It would take each tweet as it is sent, and classify it as either trolling or not. Ideally, uers would be able to mark specific tweets as trolling as well, the way you can with spam emails. Eventually, such a tool would become very good at classifying trolling tweets, and you would never have to read the hateful, crazy messages again! How nice that would be.
In this repository, I have some Twitter data and some Jupyter notebooks showing how I created a small-ish training set and trained a support vector machine algorithm to classify the Tweets. It is an on-going project that I will add to regularly as I come up with new and better ways to develop this product.