In this day of age, people seem to get sucked into their own personalized vacuum chambers. Without exposure to the other side, Even for the most aware, it is rare to confront difference.
What it does
pontem aims to bridge that divide. When a user is reading a news article, pontem alerts them to articles about the same topic that are written from different perspectives.
How we built it
We have a chrome extension that is communicating with a Flask server run on AWS. The active url is sent to the processor and is parsed for keywords. The main entities are extracted and similar articles from recent history are found on Google news using these keywords. Each article is then scored with the sentiment around the keywords. The furthest articles are then returned.
Challenges we ran into
Even though some of us had experience with Chrome Extensions, it was still quite hard to adapt to the paradigm and the unique communication flow associated with the system. One of our team members was working remote and went AWOL without contributing a single line of code. He was not the real MVP (Jaime might have been). Also we had to fix the bugs written by BST.
Accomplishments that we're proud of
What we learned
What's next for pontem
We want to write our own bespoke keyword generator and article finder. It would also be good if we trained our own transformers to run a specific keyword sentiment analysis.