This is the challenge for stock price prediction by Siraj on Youtube.
For the challenge, I used the finance data of Volkswagen with period of a year starting from Jan 4 2018. You can find it here Data. I scaled the data into minimum and maximum values of 0 to 1. Next, I retrieved the latest tweets about Volkswagen and identified them to be positive before moving on. Also, I chose to go with two layers of lstm, with last 30 records as timesteps.
keras, sklearn, numpy, pandas, tweepy, nltk, textblob and matplotlib for plotting
Just run
It is predicting based on the data, not built to forecast yet.