Veritas

Inspiration

Inspired by the results of our last project of how social media creates political echo chambers. We wanted to analyze a different side of this topic and make use of more powerful machine learning models

What it does

Our project uses a recurrent neural network to analyze text patterns to asses how likely a inputted link to a news article contains misinformation/skewed reportings.

How we built it

Using react js for the frontend, flask for the backend, various ai libraries for our model

Challenges we ran into

Navigating and connecting django backend with the react frontend, familiarizing ourselves with react js, and making the model compatible with a wide variety of sources

Accomplishments that we're proud of

Successfully deployed the frontend and backend, connected them, as well as trained the machine learning model with 99% accuracy on the data we trained it with.

What we learned

We learned how to effectively implement a frontend and backend website while exchanging data between them and executing a python script at the same time

What's next for Veritas

If we had more time, we would have tried various other training models, and possibly have scraped for a greater amount of non-political articles to expand the scope of the model. We would also like to retry to implement the backend in django.