We were inspired by the recent surge in crypto currency as well as our general interest in crypto. We wanted to have a place where we could see the general user sentiment towards certain coins, and possibly use the data later in a machine learning model.

We build this system to scrape Reddit (and other platforms later) using PRAW, then after getting user submissions passing them to the google cloud entity sentiment analysis. This then returns the sentiment towards entities in user posts. We filter out for crypto currency sentiments, and run some data aggregation on it, this is then uploaded to our database, and displayed on our website

We are overall happy we came out with a functional site which more or less includes all of the base features we wanted. We would have liked to include twitter, but sadly we did not have verified developer accounts

Next we would like to start feeding this information to a machine learning model and see if this could be a functional predictor.

Tech Stack

Language: Python3

Front End: Dash + Plotly + CSS + HTML

Deploy: Heroku

Database: Redis

Share this project: