newso

Inspiration

What inspired me to build up this project is to avoid another possible abuse of power in misinformation. By educating the public as opposed to misinforming them for short-term personal gain, we could bring back trust in communities and avoid antipathy, ignorance and resentment and possibly bridge the divide.

What it does

Newso is a website that performs both text summarization of news articles and can determine if a news article is reliable or not. The website works by allowing the user to input a URL (the url of the news article), and can then concisely summarize all of the content of that news article into a few sentences, and can determine if it is reliable or not. All of this is made possible through machine learning algorithms and python libraries such nltk (used to extract text) and scipy (contains the necessary ML algorithms to determine if a news article is reliable).

How I built it

I used python, and other libraries to build machine learning algorithms and train them appropriately in order to determine if a news source is reliable. This is of course a prototype but there is a lot of potential to further develop this project to better serve it's intended purposes

Challenges I ran into

Had difficulty deploying to Heroku but it was a pretty learning experience, not to mention I had to read up on some resources in demystifying machine learning concepts

Accomplishments that I'm proud of

The I managed to built the website

What I learned

Learned more about Python

What's next for newso

Build more sophisticated web crawlers