Inspiration

News sources are becoming increasingly biased. It's hard to know what sources to trust and when they can be trusted.

What it does

Polarized.news classifies articles with our machine learning models. A bias value is generated and presented to the user alongside related articles with different biases.

How we built it

Python

Challenges we ran into

Data collection and sourcing was one challenge. The major challenge we faced was determining the ground truth for political bias.

Accomplishments that we're proud of

We used pew research in combination with machine learning to classify articles based on their political bias

What we learned

Occasionally, news outlets will publish articles with different political leanings than what is typically expected for them

What's next for polarized.news

Mobile app, more and better data collection, CNN for NLP and sentiment classification.

Built With

Share this project:

Updates