Twitter is a great source of text data. Machine learning helps to use it for stock predictive modeling. Current repository aggregates python scripts aimed at gathering data by scraping twitter, Natural Language Processing, and generating features using sentiment analysis. Such methods were used in the project ICvsML (repository)
Look at the developed useful functions in .py script.
UPD: A couple of months ago Twitter locked the snscrape ingoing requests. Look at the details
good news: A Twitter employee says it's temporary: "this is a temporary restriction, we will re-enable logged out twitter access in the near future"