Overview

Senti is a machine learning-based sentiment analysis tool designed to classify sentiment in news for a given topic or company. Users can enter a specific company and see how they performed in the news and compare it to the stock price graph in the same period.

Senti is built with Django and React.js. The project uses Python NLTK library for sentiment analysis.

Setup

1. clone the project
2. create an account for NYT archive api
3. create an api key:
   1. go to account -> apps
   2. create an app and enable archive api
   3. get your api key
4. in the project folder create .env file
5. add the following lines and paste your key:

   KEY_NYT='{your key}'
   DJANGO_SECRET_KEY='{our django key}'
   

6. run the following commands (on Windows):

   python -m venv venv
   venv\Scripts\activate.bat
   pip install -r requirements.txt
   python manage.py migrate
   

7. open python terminal and run:

   >>> import nltk
   >>> nltk.download('all')
   

8. go to frontend/ and run:

   npm install --save --legacy-peer-deps
   npm run dev
   

9. run the script to download articles

   python manage.py runscript download_nyt_articles
   

10. run the server

python manage.py runserver