Sentiment Analysis Project

This project facilitates the calculation of sentiment values based on comments through the utilization of NLTK, Transformers, VADER, and pre-trained models such as RoBERTa. It generates a dictionary that includes positive, negative, neutral, and compound scores. Simply input the comments into a CSV file, and you can then compare the sentiment values produced by VADER and RoBERTa. Any alteration in sentiment values will trigger a mailing system, providing real-time updates via email through the implementation of Watchdogs and SMTP.

To enhance functionality, I have developed a REST API using Django, enabling the addition of comments to the CSV file and retrieval of sentiment values. Additionally, to optimize accessibility, a tunneling network has been established using Cloudflare, allowing access to the API from any location worldwide.

Installation

Step 1)

  Install Python 3.10.0
  Install git bash
  Install Visual Studio Code

Step 2)

  git clone xyz.com

Step 3)

  cd Sentiment_Analysis_project
  pip install -r requirements.txt

Running Via Terminal

  Step 1) Open start_auto_mailer.bat file in vs code

  Step 2) Change the directory of the start_auto_mailer.bat according to your path.

  Step 3) Run start_auto_mailer.bat

  Step 4) Open Main_code folder using vs code

  Step 5) Navigate to nltk_trial.py and change the dir path with the Main_code folder path.

  Step 6) Run nltk_trial.py

Running Via Gui

  Step 1) Open Main_code folder using vs code

  Step 2) Navigate to gui_v2.py and change the dir path with the Main_code folder path.

  Step 3) Run gui_v2.py

Running API Locally

  Step 1) Open sentiment_api folder

  Step 2) Run py manage.py runserver 0.0.0.0:3000

  Step 3) This will run your api on port 3000 of localhost

API Reference

Post id, comments, brandname

  POST http://127.0.0.1:3000/sentimentAnalysisApi/image/

Parameter	Type	Description
image0	image file	Required. Select the image
option	int	Required. 1) Get Sentiment Value of A Brand 2) Create new brand 3) Add id and comments
selected_csv_file	str	Required. Enter the csv file name
brandName	str	Required. Enter the name of the brand you want to create
id_value	int	Required. Enter the csv id's
comments_value	str	Required. Enter the comment on brand

Tunnelling API Using Cloud Flared

  Step 1) Open sentiment_api folder

  Step 2) Run py manage.py runserver 0.0.0.0:3000

  Step 3) Run cloudflared tunnel --url http://127.0.0.1:3000

  Step 4) You will get a temporary endpoint link from cloudflared and through this link you can access the api from anywhere around the world

Screenshots For Running Via Terminal

Screenshots For Running Via GUI

Screenshots For Running Via API