Quantopian Algorithm using Trump Tweets data. First I use a tweepy cursor to grab a page of tweets by theRealDonaldTrump. Then I score the sentiment of each tweet using textblob (shout out to TextBlob). Then I summarize daily counts of his tweets in a pandas DataFrame, calculate a 10-day rolling average count, and calculate the daily average sentiment score. Finally I shifted the df by 1; yesterday's tweets will be today's data.
I saved the data.csv file to dropbox, and used Quantopian's fetcher to grab the data. Then I tested a trading strategy based on the hypothesis that Trump's tweets with abs(polarity) > 0.2 will cause increased volatility in the market. To test the basic idea I just selected a VIX etf to trade +/- 100% of my portfolio.