More about tweets

For every stock on stocktwits, there is a live feed of tweets made by users of any kind. An example of a user tweeting about AAPL can be seen below.
Ideal tweet
When a user tweets, they have the option of adding a bullish or bearish tag to their tweet. It's assumed a tweet gives us the sentiment a user has about a stock in the near future. In the image above, this is an example of an ideal tweet where the user gives a bullish opinion about $AAPL and gives reasoning. However, while this is the primary sentiment from this user, there are many key unknowns associated this prediction. Some include: for how long they think it will be bullish, how bullish they are, how reliable this user is at predicting (big).
Non-ideal tweet
Another problem with tweet based feeds is having users that provide no useful information. This tweet doesn't have a sentiment associated with and doesn't give much reasoning or context to their sentiment. When you scroll through stocktwits, many times this is the case with tweets and can be hard at times to find any useful information. One major goal with this project is to effectively filter out unhelpful tweets and quickly gain sentiment about a stock without sifting through hundreds of these tweets.