Inspiration

Although the cryptocurrency market is not as fast paced as the stock market, but nonetheless it's very dynamic. If an investor could even save a couple of seconds on a correct analysis of a cryptocurrency, that could end up have him successfully make a trade without losing the opportunity like price changing too much which renders the initial analysis obsolete.

What it does

The bot automatically detects cryptocurrency's intermarket price discrepancies for your cryptocurrency portfolio and accounts for the respective crypto's market sentimental analysis to assess whether to carry an arbitrage trade.

How we built it

We used Python to retrieve and send data to cryptocurrency exchanges - Bittrex and OKEX. The cryptocurrency trading bot models popular arbitrage strategies including convergence arbitrage to transfer between markets, and simple arbitrage within a market between trading pairs. We also used sentiment analysis in Python and used Tweepy and NewsAPI to gauge market sentiment using news articles, twitter hashtags, and tweets to predict the direction the cryptocurrency market will go.

Challenges we ran into

The APIs of the cryptocurrency's exchanges' documentation were unfortunately very poorly worded leaded to ambiguity and confusion even for mentors. Dealing with HMAC and base64 encryption for the secret keys of the exchanges' APIs was very difficult due to its archaic system of authenticating users.

Accomplishments that we're proud of

Although we didn't finish the whole project, we managed to pull through most of the challenges we faced to a certain extent tackling it head-on. We didn't quit. We have a product that each member of our team is very proud to present it to you.

What we learned

When we started off, we didn't even know what "Arbitrage" meant but now after a mere couple of hours, we know so much more about financial trading techniques than what we started off with. Also after our project ideation stage, we didn't even know how to build the bot at all. We were able to learn what in essence it is to be a developer and simply looking all over google to learn and ending up with a great product.

What's next for Arbitrage bot

We need the bot to assess more exchanges and add more cryptos in our portfolio. Also, the bot should carry out a much more rigorous market sentimental analysis through employing Machine Learning. News and analysts' market research reports can be assessed through Text Recognition and Natural Language Processing. Another future step is to create a web portal for users to view their trades, choose the arbitrage amount, toggle the bot, and export data.

Built With

  • bittrex-ocex-tweepy-newsapi-python
Share this project:

Updates