CryptoSight

Inspiration

All of us have keen interests in the computer science and finance sector, which perfectly aligned with the blockchain theme. Since some of us are avid investors along with the theme, we decided to make a website that gives the user insight into a specific cryptocurrency performance.

What it does

CryptoSight allows the user to pick between 5 different algorithms (LSTM Machine Learning simple moving average, Simple moving average, momentum trading, reverse trading, pairs trading), and 5 different crypto-currencies (Bitcoin, Ethereum, Tether, Ripple, and Binance), and the website displays the crypto's price sell ratio, Sharpe ratio, total profit, total loss, and a graph of the crypto throughout the time frame.

How we built it

We coded algorithms in QuantConnect to run simulations on different crypto-currencies, and used the QuantConnect API to send data to display on our website, which we used reactJS to create.

Challenges we ran into

We had a hard time getting the QuantConnect API to send data through to the website, and we even spent a good amount of time trying another way. However, we managed to figure it out a few hours before submitting.

Accomplishments that we're proud of

Our success with QuantConnect after digging through tons of forums & documentation is what we are most proud of with this project. We also like the minimalistic design of the website we have.

What we learned

We learned about the dangers poorly built algorithms have. Some had too much invested in every trade so it is too risky to be safe in the long term. We saw the effects of the bad algorithms when we ran them in different periods.

What's next for CryptoSight

Adding more cryptos, changing the period for testing of the algorithm, and allowing the algorithms to be modified.

Built With

• flask
• javascript
• machine-learning
• python
• quantconnect
• react
• tradingalgorithms