Almost everybody consumes media everyday. Especially people not familiar with online services, actual facts on certain topics or news sources in general might tend to be more vulnerable to Fake News. This counts especially for young children entering the "media world" for the first time or naive people

What it does

Our App offers two possible use cases:

  1. Check a Link to an article or specific text by manual insertion to the app. For better ease of use that can also be done out of other Apps where the user consumes the media. Let's say someone reads an article on Facebook, she might just mark it and "share" it with awareNews immediately seeing the probability of it being fake news
  2. The "news alert mode" - if turned on - analyzes the articles a user is consuming and only gives a warning if a certain threshold is passed, i.e. it is 75% certain to be fake news.

Lastly the App provides the user with feedback on why it is considered to be Fake News by tags. Users therefore learn how to analyze media and detect unreliable information.

How I built it

Using a Classifier Data set of appr. 5,000 labelled data and almost 30,000 data points of training data our Machine Learning algorithm is using logistic regression to get a percentage of how probable it is to be "Fake News". This is implemented in Python. The backend uses Java to process the information. The App itself ist programmed in Swift and data is exchanged using Firebase.

Challenges we ran into

Finding the right data set is quite a challenge - especially having in mind that it is an ambiguous topic. In the end there was a usable data set we found on Keggle.

Accomplishments that I'm proud of

An easy to use solution to one of the most important topics of our time. Especially the ability to run it in the background should be a great feature to enhance an handy and every day use.

What's next for awareNews

Implementation of a community feature to engage users more

Share this project: