Trading stocks on Twitter data
We have created an algorithm that trades based on Twitter data. We spent a year and iterated through five algorithms to reach this point. We have beat the market with this algorithm, not only using backtests, but using our own hard earned money.
In one month we made almost 10% returns (11/19/17 - 12/19/17) This is our own hard earned money. We trust it because we have seen its results.
During the same time, the S&P 500 made only 4.2%.
We collect data from twitter and predict on 8 companies, listed below. Each day the algorithm predicts which stocks will go up, and which will go down and trade automatically based on that. We can choose any company we like, but we just have these 8 for simplicity and because the twitter API is not free. This works because sentiment analysis extends beyond technical analysis to poll the internet's opinion of companies.
Our backtests of 8 months also show that we beat the market by over 2x.
The figure below shows returns of a mutual fund like strategy with the same 8 companies in a buy-hold strategy, compared to the benchmark. We also beat that, showing our returns were not soley due to our stock picks.
Cofounder & CEO
Cofounder & CTO