/Polarity

Twitter sentiment analysis website

Primary LanguageJupyter Notebook

This project was bootstrapped with Create React App.

Polarity

A website that displays the percent positivity the top tweets that contain any hashtag that's been searched

Tech Stack:

ML- NumPy Pandas scikit-learn

Server- Flask

Frontend- React

How it Works

Tweets- The tweets are gotten in real time using the library snscrape and filtering the tweets by the hashtag entered by the user.
Model- The model has been built using the Logistic Regression algorithm and trained on the Sentiment140 dataset, currently it has reached an accuracy of 78%

To set up a dev environment for this repo-

Fork this repo and navigate to the client directory, in the client directory, you can run:

npm start

To fire up the frontend

Server

To start the server, go to the server directory and run:

flask run

the website can then be accessed at localhost:3000