I have built a LSTM network that is able to predict the price of stocks given 30 days of historical data. The information of the stock price prediction, alongside other metrics such as the EPS are displayed in this interactive dashboard.
The environment folder contains the conda environment used to develop this project.
- Improve performance by creating a callback that serves all the components. Currently, a unique function is used for all the different interactive components, which means that actions are duplicated and the load time takes longer than necessary.
- Add additional metrics to the dashboard. The metrics displayed at the moment are: graph showing historical and prediction price, the prediction price, and the EPS (Estimated and Reported, as well as the surprise)
- Improve the model used to predict stock prices. The model was built on the S&P 500, meaning that this is not very accurate at the moment when analysing individual stocks. Additionally, we will need to implement a way to update the model automatically.
See the open issues for a full list of proposed features (and known issues).
Edwin Jaico-Berg - @ejaicoberg - e.j.berg@outlook.com
Project Link: https://github.com/EdwinJaico-Berg/stock_predcitor
Image: 'financial report' by Nuzulis Syafaah from Noun Project)