Stock Price Predictor

A project to start learning about machine learning in Javascript. Utilizing the Brain.js library and the stock prices API provided by Alphavantage.

The aim of the server will take a stock symbol, pull the last 100 days of stock data from Alphavantage then make a prediction on what the stock price will be today! We also save each prediction in a database to make checking if the predictions came true.

I also wanted to use this project as an opportunity to work with Server Side Rendering of a client app. I chose Pug as a template engine and run all the basic functionality of the client though vanilla Javascript. All the client side files can be found in the public and views folders.

Prerequisites

• A MongoDB instance running

Getting Started

• Clone the repo
• Run npm i
• Add a .env file and add:

JWT_SECRET=<anything you want>
ALPHA_VANTAGE_KEY=<API key from Alphavantage>

• npm run start in the terminal to get the project running on port:8080

• Then visit http://localhost:8080/ in your browser to access the app.

Calling the API

Currently the only route that can be called on the api is:

http://localhost:8080/api/stocks/get-stock-prediction/:stock

params:
stock - needs to be a valid stock symbol e.g AAPL or MSFT

You will then get a response back that looks like this:

{
    "success": true,
    "message": "Prediction made successfully!",
    "data": [
        {
            "open": 196.95269326686858,
            "high": 200.28645069122314,
            "low": 195.04503690719602,
            "close": 195.04503690719602
        }
    ]
}