Stock Predictor

Does automated trading actually work?

Ever wondered if there’s a correlation between company news and stock performance? That’s where Stock Predictor comes in—an open-source tool I built that monitors Reuters via Twitter, scanning articles for sentiment signals about stocks you care about.

The program analyzes article sentiment and suggests a direction. For example, if Reuters publishes a leak about Apple launching the iWatch and what a huge success it will become, the app will calculate a positive sentiment and mark it closer to 1. Higher scores trigger buy recommendations; lower scores suggest shorting.

You can find the project here.

Give it a spin, fork it and send me a pull request if you think something can be improved here. My vision for this a poor man’s high frequency trading app. All that this would be missing is a broker who is willing to give you electronic access actually trade electronically trade on a market. This type of access is called leased lines and firms spend quite a bit to have this kind of access. At the very least, you can simulate your paper profits of stocks you are interested in and who doesn’t like doing that!

Leave a comment