Stock Helper (won 2nd in a hackathon)

An app made to help investors/traders by providing stock predictions and transcript summarization

Features

• Stock Prediction models: The website features a stock prediction model which can help in predicting future stock prices.
• Summarized Transcripts: The Website provides a summarized transcript of the quaterly fiscal meeting.
• User-Friendly Interface: The website features a clean and intuitive user interface that is easy to navigate, with clear menus, search functionality, and customizable settings. It is designed to provide a seamless user experience, whether users are beginners or experienced investors/traders.

Technologies Used

• Front-end: ReactJS
• Back-end: Python, Django, Pytorch
• APIs: Stock market data APIs for real-time stock data and financial information

Model Details

• LSTM model
• Price Prediction Model is Trained on 80% of the data.
• Unique model for each stock
• Tech used: Pytorch

Key points

• backend/stocks/price/make_model.py used to make models
• backend/stocks/price/predict_with_model.py used to predict price
• Data was preprocessed using the file Preprocess_Data.py
• Transcripts were summarized using Summarize_Transcript.py

Steps to Run

1) Locally Hosting Backend

In "Backend" Directory

• Create a virtual environment if needed using:

   py -m venv .venv
  

• Activate virtual environment.

    .venv/Scripts/activate.bat //In CMD  
    .venv/Scripts/Activate.ps1 //In Powershell
  

• install requirements

     pip install -r requirements.txt
  

• Running backend server change directory to \stocks then do

  python manage.py runserver
  

2) Locally Hosting frontend

• In "Frontend" Directory do

•    npm i 
     npm start
  

• Webapp should be running at localhost:3000

3) Select a stock by searching for its symbol on the search bar on the top left