Python Stock Market Analysis

The process

1. Get historical data
   • Stock market open/close
   • Archival finance news (i.e. New York Times finance section)
2. Create mechanisms to gather current data on a daily basis
3. Perform sentiment analysis against news data to determine positive v. negative coverage.
   • To keep it simple for now (and avoid the need for manual classification and training), I will create one list of positive terms, and one for negative terms.
   • If more positive terms exist between open and close, the days news will be considered positive, and vice versa.
4. Determine if there is a correlation between the sentiment of the news stories and the market's behaviour.
5. Use the data collected to train/test a machine learning algorithm.
6. Predict future stock market behaviour based on the sentiment analysis of the previous days news.