My new project

have a try at this link guys, my new app for predicting stocks for the next 30 days using stacked lstm model. Please do let me know your valuable comments and on the aspects that i can improve on my next project.

Note: The app is created only for educational purposes and not for commercial usage.

1 Like

Here are my quick and unembellished comments :wink:

  • layout is too small, especially when showing charts
  • No error message if user enters an unknown ticker symbol
  • case-sensitive is annoying, just use .upper() string method
  • Predictions as text??? We also want to see a graph!!!
  • If one is honest, one would now also have to show how the price predictions match the real future, i.e. supplement a proper backtesting :sunglasses:
  • About > Procedure > Steps Involved is not formatted well, use a markdown block and proper markdown synatx
  • your whole application is embedded in a try...except block, this is maybe convenient, but bad practice and covers errors if something goes wrong
  • your config.toml is in the wrong place
  • cleanup your requirements.txt file
  • I won’t write anything about the usefulness and success of stock price forecasts with the help of Machine/Deep Learning, that’s too controversial… :zipper_mouth_face:

Like mentioned in the beginning of the app that this is just for educational purpose, which means its not for commercial usage. I feel this ain’t controversial unless and until sentimental Analysis is involved.

No one can give a prediction of 90% when comes to stock market. Factors affect the price for every second. So I hope people don’t tend to take this application seriously for investing and make this controversial. I have created this application to just establish myself that I can solve complex real world problems, being this as my first project in streamlit.

Thank you for mentioning your discrepancies. It will be corrected very soon.