Analysis of Newspaper Propaganda Biases

We present a system which visualizes newspaper biases on the basis of propaganda techniques using PRTA API. The system is an umbrella containing two applications, a JS frontend powered by d3.js and a flask backend.

The JS frontend can be hosted on a simple live server. We recommend a VS code live server or a live python http server using

python3 -m http.server [PORT NUMBER]

For BE, the following packages need to be installed:

• flask
• flask-cors
• numpy
• pandas
• scikit-learn
• scipy
• spacy
• nltk

Spacy and NLTK require further packages which can be installed by running

python -m spacy install en_core_web_sm
python -m nltk install stopwords

To run BE, run

flask run

The backend abstracts the following APIs

• /<search_criteria>
• /article/<article_id>

Files

Kaggle Dataset: https://www.kaggle.com/snapcrack/all-the-news

Original Kaggle data of the "All the News" dataset are the files: articles1.csv, articles2.csv and articles3.csv

Install the data files and store them in a directory named 'archive' in the project directory. With this, you are all set!

Cheers!