This project is an attempt at performing a simple periodic sentiment analysis of given stock symbol on reddit.
- FastAPI
A scheme of the whole project architecture can be found at on figma
In short, the main process are:
- Every 4 hours, a script scrape the latest posts and comments associated with a stock
- The script send the data to a sentiment analysis API that analysis the text
- The sentiment analysis API send the data to store it in a NoSQL database
- A static web interface that fetch the data through the FastAPI DB
The reddit posts are fetched from the API with the format http://www.reddit.com/r/{subreddit}/new.json?
Other useful options that can be used in the queries are:
limit={limit_number}(note that the upper limit is 100)sort={sorting_option}
In the building of the project, the following ressources are providing usefull informations: