As a team, we are all very interested in current events and how an informed public can shape political outcomes. However, there are few programs out there that actually can determine how mainstreamed an article, news source, or any text file is. Our goal was to create a program that could output an objective calculation of how biased an article is. The most difficult part of our challenge by far was how to objectify something subjective. We had to identify patterns in new sources that are commonly associated with bias, and formulated algorithms that are correlated with those patterns. Studying and experimenting with multiple articles provided a better and clearer output.
In the future, we hope to use machine learning to enhance our algorithm. We would notice patterns in the clearly biased articles, and our code would then give us suggestions to modify, enhance, and better our code. As the number of input articles increases linearly, the betterment of the program exponentially increases, since its betterment would detect more specific elements in many cases.