I was mainly playing around with algorithmic crypto trading. I created a model which takes in historical data on a crypto (mainly tested on Bitcoin USDT) and calculated "indicators" of these data, and then makes decisions on whether to buy/sell/do nothing every hour.
Both files of code are only runnable on the QuantConnect environment (they have their own libraries and such):
https://www.quantconnect.com/
main.py is for running model on historical data and testing its performance.
research.ipynb is for investigating different indicators and seeing how good they could be for the model.
Example run (one of the better results - most of mine ended up having losses lol):
