Stock Market predictions with Prophet and FastAPI

Train prophet models and run stock market predictions for a given ticker

Details

• Dev dependencies
  • FastAPI: https://fastapi.tiangolo.com
  • Docker: https://docs.docker.com/
  • SQLAlchemy-Core: https://docs.sqlalchemy.org/en/14/core
  • Databases (Async SQLAlchemy Core queries): https://github.com/encode/databases
  • fbprophet (Time Series forecasting): https://facebook.github.io/prophet
  • yfinance (market data downloader): https://github.com/ranaroussi/yfinance
• Testing:
  • pytest: https://docs.pytest.org/en/stable
  • pytest-cov: https://github.com/pytest-dev/pytest-cov
  • Codecov: https://docs.codecov.io
• Linting and Formatting:
  • black: https://github.com/psf/black
  • flake8: https://flake8.pycqa.org/en/latest
  • isort: https://pycqa.github.io/isort
• Continuous Integration by Github Actions

Actions

• Train the model for a ticker
• Check the weekly prediction for a previously trained ticker
• Check the weekly prediction for all the trained tickers

CRUD

• Create prediction for that ticker
• Read results for all tickers or for individual ones
• Updating the info for a ticker should be done by creating a new one due to its time-sensitiveness
• Delete a specific ticker

Future additions

• Authentication
• Extra information fields based on third-party content
• Add plots and images if available from source
• Input training timeframe
• User Interface