4Thought

4Thought is a web application that takes the url link to any article before determining if the article is biased while at the same time giving the user related resources.

Comment 1

Inspiration

Meta's challenge piqued our interest due to our interest in current political topics. We were also eager for the opportunity to use react for Meta's challenge. The things react has done for development frameworks is an inspiration in itself.

What it does

Our application takes in the URL of political news articles and determines if the article is biased. We have three categories of bias: Left-Leaning, Right-Leaning, and Unbiased. When presented with your article's bias there will be three similarly biased articles.

How we built it

We used a combination of Flask and React to create an interactive dynamic experience for users. The backend database is made using SQLite. The flask API is the bridge between that database and the front-end server. The front-end React UI interacts with the API from its own server making the cloud-based solutions independent from one another. The servers we used were provided by rep-lit.

Challenges we ran into

It was challenging creating the scraper in a way that the keywords could easily be inserted and extracted from the database, along with the logic for how the scraper would accurately collect data from websites. Applying Cross-Origin Resource Sharing (CORS), and properly linking the frontend to the backend caused issues, as well as the servers, are located at different IP addresses.

Accomplishments that we're proud of

We are really proud of how well the application functions. It is not only easy to use, but given the time and resources, it could be incredibly accurate.

What we learned

This experience has helped us learn how to work under pressure as a team while establishing a good workflow remotely.

What's next for 4Thought

We have many future plans to keep this project running in the long term. We would like to implement machine learning to help better detect biases between articles. We would like to add more computing power and storage to process more articles to check biases against. Lastly, if we could add a system that deletes some of the older recommended articles while adding the searched articles to keep the system more up-to-date as they are stored with their classification anyways.

Share this project:

Updates