News Sentiment Anlayzer

Inspiration

We noticed that fake news has been an enormous problem in the past year and decided to find a way to determine the amount of bias in a news article.

What it does

We built a website where the user can post any text into a textbox in order to check its bias. Because the training data for the deep learning algorithm was based on news articles, it functions best with statements from news articles.

How we built it

We built the front-end of this project in html, css, and javascript while the backend was written in python using paddlepaddle for some of the algorithms. All of the components integrated well in the Django framework. In this way, we could match the backend with the frontend and also allow the project to communicate with docker to do some of the machine learning processing.

Challenges we ran into

Our biggest challenges were installing the appropriate software and using PaddlePaddle's API in our own way. Also, fatigue hit us hard towards the end. In fact, as I'm writing this, my teammate is passed out on the ground beside me.

Accomplishments that we're proud of

Stayed up all night. Parker didn't throw up.

What we learned

You don't spend very much time coding at hackathons, surprisingly enough. We spent much more time installing software and reading API and documentation.

What's next for News Sentiment Anlayzer

We need to add more features to vectorize. Also, we'd like to try a different machine learning algorithm and expand the data sets.

Built With

• docker
• html5
• javascript
• paddlepaddle
• python
• sass