We've known about confirmation bias for centuries, yet its stranglehold on society seems to only get worse by the day. Our tendency to misinterpret ambiguous articles to support our beliefs is positively correlated with the global increase in social media usage. And how we congregate in groups doesn't help the issue. Simply put: liberals talk to liberals; conservatives talk to conservatives, but no one actually talks to anyone.

Hack News changes that.

Hack News utilizes Indico's API and machine learning algorithms to examine your social media and determine what political preferences and biases you possess. The application doesn't view political preferences as binary; it recognizes the myriad of opinions and perspectives regarding fiscal and social issues. Hack News then takes that information to generate a custom Google search which outputs news articles with opinions contrary to your own.

Whether someone identifies as a Democrat, Republican, Libertarian, or anything in between, we believe that it's a fundamentally good thing for people to expose themselves to a diverse set of opinions. Google's and Facebook's tailored search engines are great for convenience but bad for genuinely exchanging ideas. Our custom engine intentionally makes the user read from qualified authors who might not agree with her beliefs. That may make her uncomfortable and cause some cognitive dissonance. But spending time to understand different beliefs and exchange ideas is what ultimately drives our democracy forward.

What is it?

We developed Hack News using HTML, CSS, and JavaScript in order to create a visually appealing webpage that returns news articles that are outside of your political bubble. We created a node.js server to run it all, and utilized jQuery in order to efficiently call Indico to utilize sentiment analysis, characterizing the headlines of chosen pages of being either conservative, green, democratic, or libertarian. In order to do this we used a binary search of a handful of news organizations, cross-applied with the pages the user likes. We then developed an algorithm to go through all the political posts of the pages that the user liked and use sentiment analysis to determine the political bias of the post. This would create a customized search engine for the user.

What did we learn?

This was our first time using APIs and custom search engines. We learned about using APIs through the use of Facebook and Indico API.

Share this project: