AbhijithGanesh / StockSage

My stock analysis project using LSTM and SARIMA. This is a test project and it is not financial advise.

Stock Price Prediction Project

Introduction

This project aims to predict the stock price of Apple Inc. (AAPL) using data from Yahoo Finance via the Rapid API. The prediction models are based on SARIMA (Seasonal Autoregressive Integrated Moving Average) and LSTM (Long Short-Term Memory) implemented using TensorFlow. The project also utilizes Grafana for data visualization, PostgreSQL for data storage, FastAPI for REST endpoints, and Docker for containerization.

Disclaimers

This is just a sample analysis and it is not financial analysis for Apple Inc or any stock. I've tested out both SARIMA and LSTM for this project and have stored the weights of LSTM

Prerequisites

• Python 3
• Docker
• Grafana
• PostgreSQL
• FastAPI
• TensorFlow
• Rapid API Key

How to get started

• Clone the repository

git clone https://github.com/AbhijithGanesh/StockSage.git

• Set up your environment:

Refer the .env.example

POSTGRES_USER=''
POSTGRES_PASSWORD=''
POSTGRES_DB=''
DB_HOST=''
RAPID_API_KEY=''

• Installing python dependencies

pip install -r requirements.txt

• Spin up the Docker Instance:

docker compose up

• Run the FastAPI server:

uvicorn app:app

Acknowledgements

• Thanks to Yahoo Finance and Rapid API for providing the stock price data.
• The SARIMA and LSTM models are built using TensorFlow.
• Grafana and PostgreSQL help with data visualization and storage.
• FastAPI provides the RESTful API for accessing predictions.
• Docker makes the project containerized and deployable.

Tech Stack

• Data Fetching:

  • Yahoo Finance
  • Rapid API

• Data Storage:

  • PostgreSQL
  • (TimeScaleDB will be implemented in the future)

• Data Visualization:

  • Grafana

• REST Endpoint:

  • FastAPI

• Machine Learning Models:

  • SARIMA
  • LSTM (Long Short-Term Memory) using TensorFlow

• Containerization:

  • Docker

License

This project is licensed under the GPL. Feel free to re-use it😉

Sample Visualization