Peer Review

[zm57]

Things that I like:

1. I love this topic because I truly believe that sentiment plays a huge role in the stock market and sometimes it is the dominating factor the moves the market.
2. I like how you are employing NLP to target this problem. A very important part of this problem is to successfully analyze the content of the tweets and the available market information.
3. I like how you pick DJIA as an indicator and representation of the market. I think it is very representative.
   Things that need to improve:
4. I am slightly worried that simply using YAHOO tweets and financial news are not going to give you a comprehensive representation of the sentiment in the market. Something I found out is that people who actually tweet on those forum actually insanely positive and long the specific indices or stocks that they tweet about. The languages and tweets on those forum can sometimes be very biased. I also want to introduce you another website for stock tweets: StockTwits.
5. You probably want use other numerical factors in predicting your result. For example, VIX, the volatility index is something that reflects investor's sentiment.
6. If possible, probably try to test out your thesis in the other markets. Even though DAJI is a good representation of the US market, it is not neccessary a good representation of other markets.