Most discussed stocks of last week!
How does it work: I use Python to stream (Reddit and Twitter API) and store tweets and comments containing mentions about stocks. I added sentiment analysis to the findings so as to check the general sentiment (whether what is being talked about the stock is positive or negative).
What is sentiment: It's calculated using natural language processing with a modified version of vadersentiment. You get pre-trained models which classify whether a given statement is positive or negative. I have modified it to analyse financial comments also ( eg. Calls, puts, moon, gain, loss etc). It then classifies the statement between -100% and +100% depending on if it's negative or positive respectively. The numbers you are seeing here is the average across all the comments.
here is a sample of the sentiment analysis I posted when Boeing's engine failed on UAL flight.
Is it 100% accurate? No! The program misses out of nuanced sarcasm. But when you are analysing 100's of comments the issue is not very prominent enough significantly change the results!
Is the program public: Yes. I have shared the GitHub repo for the code (pinned post in my profile, if mods allow I will update the GitHub link here). You can use it yourselves to track the stocks you like.