Sentiment Analysis on Social Media

This repository contains the codes and models used for creation and deployment of a Sentiment Analysis project based on text.

How does it work

• It uses Natural Language Processing to determine the Sentiment of a given text.
• It was created in Python using libraries which are used in Machine Learning projects.
• The different models saved in the ".sav" format were trained using datasets of more than a million tweets and thousnads of reddit comments.
• The final application can predict the sentiment of a given text as:
  • Positive :)
  • Negative :(
  • Neutral :|
• The files saved on the main page of the repository are used by Heroku to access the models and other relevant code.

NOTE

If the Application on Heroku doesn't seem to be working, you can check out this link for images of the deployment:

https://drive.google.com/drive/folders/1UHbHCJYygisRgDRbGGsSbryL3C1vgpyt?usp=sharing

The above link also contains codes and datasets which are relevant to this project.

Codes and Dataset

To get more information on the codes and datasets, please read CodeInfo.md available in Notebooks and Datasets Folder. You can also open it directly using:

https://github.com/indianeagle4599/Major-Project/blob/main/Notebooks%20and%20Datasets/CodeInfo.md