Algorithmic trading continues to increase in popularity but is too complicated for your casual trader. It requires years of knowledge of stocks, investing, and programming skills. We wanted to remove these barriers and make algorithmic trading accessible.
So what did we do?
We built an direct English sentence to algorithm generator. We use natural language processing to analyze a sentence. We feed the analysis to a custom-built compiler that generates an investing algorithm. We leverage Quantopian and Zipline to backtest the generated algorithm against historical stock data. We then use various Python data visualization libraries to display this data in an intuitive and informational way. This allows user to see a preview of how their algorithm would perform in the real world.
We finally post-process and output the generated algorithm as a .py file. Write your Investopedia Stock Simulator username and password to line 10 of the file. Then simply run it with Python and it will begin trading live, under your account, on the simulator.
Accomplishments that we’re proud of:
Our entire project was very involved - we had a web app, a natural language processing, data visualization, backtesting, and live simulating. These parts had to work together flawlessly to accomplish what we wanted. Though we ran close on time, we were able to get it done.
What we learned:
We learned a lot over the weekend, everything from how to properly analyze text using natural language processing, to how to build our own compiler, and two different web technologies work together.
What’s next for Rhythm:
Though we were able to add a lot of cool features, there is a lot more to investing than we currently accommodate. We want to make our natural language processing more sophisticated, allowing for more complex investment strategies.