The Social Trader

Inspiration

In today 's economy, investing and making passive income is essential to shield against inflation, recession and financial instability. Using technology to learn more about how trading works and how we can use data to analyze the market and the impacting factors and make money, was very interesting.

What it does

It is a trading bot that resembles a market maker. It watches the order books for available instruments in the market and places sell and buy orders attempting to make small profits on each trading iteration. It monitors the social media for posts about the traded instruments, analyze the sentiment to predict risks or opportunities associated with these instruments and make trading decisions based on those predictions.

How we built it

We followed the extensive Optiver guide to build a basic trading algorithm and then enhanced it by using the LLM classification model to analyze the social feeds. We tried different combinations of parameters to improve the trading performance and minimize the losses/risks. We used Cloud9 environment provided by Optiver to develop the bot and to run it in the Optibook environment.

Challenges we ran into

We came into the challenge with no prior knowledge about trading or sentiment analysis. Understanding the mechanics in order to make the right decisions was hard.

What we learned

• The basic terms about trading and market makers.
• Sentiment Analysis Classification

What's next for The Social Trader

• Explore different classification models to get the sentiment of social posts and also correlate posts to certain instruments.
• Explore different tweaks to the algorithm to make better decisions

Built With

• facebook/bart-large-mnli
• jupyter
• python