Django web app designed to summarize and detect political bias in text
First clone the reposirory, then optionally create a virtual environemt and activate it. Next pip install all the requirements. Finally cd into the 'newsBiasSite' folder and start the django server.
git clone https://github.com/kyle-mcintyre/NewsForYou.git
python -m venv venv
./venv/Scripts/activate
cd newsBiasSite
pip install -r requirements.txt
python manage.py runserver
- The home page is where you can paste in an article to be summarized and politically scored. There is also a side box to add a comment.
- The about page describes the project from a high level.
- The comments page keeps track of all comments in chronological order (newest to oldest)
The AI Model is a support vector machine created from ~4,000 summarized political articles. Each article was scraped from far left, neutral or far right news sites and scored as either a 1, 3, or 5 accordingly. News sites were determined from https://www.allsides.com/media-bias/media-bias-ratings.
- Python
- sklearn (SVM)
- gensim (text cleansing)
- django (website and database)
- HTML
- CSS
- Javascript
- SQLite
Kyle McIntyre
GNU General Public License v3.0