/Kryptonic

A digital currency investment app which will help you decide in which coin to invest based on its history, using machine learning on it.

Primary LanguageJavaScriptMIT LicenseMIT

[PRs Welcome] [GitHub contributors] [GitHub commit activity]

Overview

  • A digital cryptocurrency investment webApp which will help you decide in which coin to invest based on its history and trends, using machine learning on it.
  • When the user enters the coin he wants to invest in and the time for which he/she wants to invest, we train a personalized Prophet model on our Microsoft Azure machine and predict net profit in the form of Kryptons💠!! We can also compare multiple coins and show the best one to invest in. Everthing happens realtime in just a matter of some seconds!!!⏱️
  • Kryptons💠: It is statistical value developed by our team which tells us the confidence of investment, Higher the number of Kryptons, Higher is your chance of a good investment!
  • Project information ℹ️

    • Project Name: Kryptonic
    • Short Project Description: A digital cryptocurrency investment app which will help you decide in which coin to invest based on its history and trends, using machine learning on it.
    • Team Name:Pandas 🐼
    • Team Members: Naman Bansal @psifrous, Swapnil Panwala @panwalas, Aniket Chowdhury @aniket-chow, Sahil Khandelwal @sahilinjaipur
    • Repository Link: https://github.com/psifrous/Kryptonic
    • Labels: Microsoft, Azure, Facebook, Machine Learning, MLH, cryptocurrency,cryptocompare, blockchain, influence the masses,webapp, Time series

    Link to the code: https://bit.ly/2Myg4su

    Platforms used 📦

    This project was developed on the Azure platform
    • WebApp module of Azure to host the website(webapp) which is available at http://mlh-jaipur.azurewebsites.net/
    • Virtual Machine to keep our ML Model running on the cloud using Flask at 13.90.227.9
    • The FrontEnd was built using Facebook's ReactJS
    • The is on Azure is being done by Node's Express Server

    Our ML algorithm is based upon Facebook's Prophet algorithm. Prophet tries to capture the seasonality in the past data and works well when the dataset is large.
    Please refer: https://www.analyticsvidhya.com/blog/2018/05/generate-accurate-forecasts-facebook-prophet-python-r/

    Below We have implemented our predict function using prophet for ETH (Etherium) to invest for the next days and in the graph you can see with the red color in the end that it has predicted the next 10 values for the next 10 days.
    Graph : Actual (blue - y) vs Predicted ( red - yhat)

    Project Synposis ⚛️

    Ever had a thought of investing money in a smart way?
    Sshhh🤫! We have a solution!

    Current Opportunity 🚀
    There are numerous ways to invest in a stock market... But! Don't you think it's differs from region to region and the market is quite saturated and common?
    So we came up with Kryptonic a unique webApp that lets you compare different cryptocurrencies and suggests a personalised cryptocurrency for a given amount of time, based upon a Machine learning model... and all of it happens real-time!

    Local Deployment

    The project can be locally deployed using nodemon index.js in node-server folder after installing the required dependencies.

    Contributors ✨

    Naman Bansal
    Naman Bansal

    Swapnil Panwalas
    Swapnil Panwala

    Aniket Chowdhury
    Aniket Chowdhury

    Sahil Khandelwal
    Sahil Khandelwal

    🔥This project was made for lnmhacks hackathon held at LNMIIT, Jaipur.