Quant Trader is a Python project which attempts to predict future stock market and cryptocurrency movement through the utilization of deep learning with long short-term memory layers. The project has three forms of predictions: closing price some time in the future, next day candlestick, and candlestick some time in the future.

This program creates datasets from Alpaca API, modifies them, and then uses them to train a neural network to predict future close prices or candlesticks.

LSTM trained with hourly closing values of Bitcoin from 2019-08-22 to 2022-08-22.

Bi-LSTM trained with hourly closing values of Bitcoin from 2020-08-22 to 2022-08-22.

Bi-LSTM trained with hourly closing values of Bitcoin from 2019-08-22 to 2022-08-22.

• Stock predictions
  • next day candlestick
  • closing price some time in the future
  • candlestick some time in the future
• Parabolic SAR and Fibonacci Retracement (needs uptrend/downtrend)
• Cryptocurrency prediction
  • next day candlestick
  • candlestick some time in the future
• Client to trade on paper account

• Alpaca API
• pandas
• numpy
• TensorFlow
• Keras
• scikit-learn
• plotly
• matplotlib
• requests
• json
• os