Inspiration
The inspiration for this application is the fact that the media has such a huge impact in our lives; so because of that we want to know what kind of "data" we are taking into our brains before we let it directly impact what we think about certain things. Having an Natural Language Processor that will tell us if some text is written in the form of pure emotion, opinion based, or straight objective facts would we quite useful. We could then know what is fake, what is fact, and what is someones opinion. This would especially come in handy when looking into news articles. Detecting fake news, fact based news, and opinion based news.
We chose the name Aristotle, due to him being the founder of the three appeals: ethos, pathos, logos.
What it does
Aristotle takes in textual data, say for example a news article, and classifies it into one of three categories: ethos (how credible is the author, and how much opinion), pathos (is the article an emotionally oriented piece with little logic), and logos (is the article purely objective). We score each of the categories, and show how we got to the conclusion. If it is Logos, we use a fact checking system to let you know if the article as a whole is factually correct, and break it down sentence by sentence if wanted, showing what are facts and what are not. If Ethos, we tell you about the author, and how credible they really are. We also tell you what kind of opinions they are showing toward what they are writing about. If Pathos, we tell you how much emotion, and what kind of emotion are in the piece.
Having this data lets us do many things, such as let people sort news by pathos (happy or sad, etc), check facts of articles, seeing how logically it really is, with sources, and see the credibility and opinions of authors.
How We built it
We used Googles Cloud NLP and Fact Checking API, along side NodeJS. We are focusing on the backend side of things, so we made a simple scrappy frontend with the time we had left in some simple HTML, CSS, and Javascript server-side rendered with Jinja and Express.
Challenges We ran into
The first major challenge we ran into was the Google Fact Checker API not having good documentation at all. Even the Google Cloud Representative present said it was pretty terrible. :( Once we figured it out, the next challenge we had was something that set us back quite a few hours. We took forever to find an error with some requests we were making, and the way it was sending it to us, once again with the Fact Checker API.
Accomplishments that I'm proud of
A working application; figuring out an awesome API with little to no documentation; making a tool that will really help everyday people with knowing what kind of things they are reading.
What We learned
Basics of NLP Google Cloud Services, such as NLP API and Fact Checker API More NodeJS REST APIs inside and out
What's next for Aristotle
We see great value in this project, and want to take it to another level. We have tons of ideas to keep this project going in a forward direction to helping people. We see ourselves creating our own NLP AI, a better Fact Checker API, and making it accessible to regular people. The concept can go a very long way.
Log in or sign up for Devpost to join the conversation.