Inspiration
There is a lot of anger in outrage in the news, because anger and outrage sell. However, the world is not as bad as news outlets would like us to think, and life would be just a little better if we could filter the negativity.
What it does
The Happy News Channel is run by Happy News Cat. Give him a URL and he will tell you how negative (or positive) the article you are about the read is!
How we built it
We used machine learning (with nltk) to train a classifier that takes in text and decides whether the content is positive or negative. We use an AWS virtual machine to host the website that uses a domain from domain.com. The user can feed a url to the website and a python script in the background will extract the text from the given webpage and pass it through the trained classifier which will tell you whether the page was positive or negative.
Challenges we ran into
The machine learning nltk stuff is pretty tough. And running the training sets was tough on the computers - need more memory :( We would have needed more training sets for better accuracy. We also had trouble finding tailored training sets. Also dealing with websites was something that was new to most of us, and getting stuff working took a little bit of fiddling.
Accomplishments that we're proud of
We got the machine learning thing working! We'd never made a website before either so that's cool
What we learned
We learnt a lot about websites and machine learning with nltk and processing natural language data.
What's next for Happy News Channel
Make it better!
Log in or sign up for Devpost to join the conversation.