-
Dashboard
-
Publisher In-depth page
Inspiration
Seeing how much headlines could differ for the exact same event by different news sources made me curious of the general bias in the online news industry, as this has become the main source of news for people, bias can be harmful and often is a sign the whole story isn't being told.
What it does
Enter a term you are interested in the news lately (example: North Korea), then enter three news sources you are interested in comparing, Sensationalyzed will use the WebHose api to fetch news articles by those publishers which feature your term, and then transfers that to Microsoft Azure's Sentiment Analysis API, which scores it on a positivity scale of 0.0<x<1.0. It then creates an average rating over every article in the past 30 days.
How I built it
Node.JS, framework is Express.JS, and the db is MongoDB, I used the Cloud9 IDE to easily combine all these and manage my project all in one space.
Challenges I ran into
Figuring out how to trim and fit all of the news data into Microsoft Azure's API became a problem, regular Node.JS and Express frustrations. Figuring out how to use Chart.JS to graphically display some of the data became confusing with the data coming from different sources.
Accomplishments that I'm proud of
Making it relatively easy to use, implementing chart.js in a sensible way, and and linking up the two APIs in a functional way.
What I learned
Sentiment analysis is really cool, however it has a lot to go as Microsoft sometimes classify's objectively positively phrases negative, and vice versa.
What's next for Sensationalyze
Deploying it online so everyone could use it.
Log in or sign up for Devpost to join the conversation.