Hi everyone, 

I wanted to share a simple algorithm that uses the mean of sentiment overtime to make trades. It longs a stock if the current sentiment is 10% or more above the average sentiment and shorts if the current sentiment is less than -1. The keywords from the Bootcamp were used as a quick input, but I recommend researching industry-specific keywords to maximize alpha.

A manual universe was used for quicker backtesting, but you can also build a dynamic universe selection model that picks "trending" stocks. Keep in mind when creating a universe for this alpha, be sure to pick stocks that have relatively consistent articles.

If anyone wants to use this algorithm as a template for further development, feel free.

 

Cheers,

Jovad