The Nexus News App: An AI-Powered Platform for Identifying Political Bias in News Articles
Team Introduction:
We are a team of 5 students from NSUT who have just completed their second year and are looking forward to enhancing their knowledge. Through this hackathon we have gained various insights on developing big scale project.
Team Members:
Pritish Mahajan (Leader)
Rujul
Harsh
Revant
Shreyaa
Event Pictures:
Problem Statement:
Tackle the challenge of discerning unbiased news in today's information-saturated environment.
Idea:
Introduction-
In today’s world of Internet and people being sharply divided among contrasting views, the media has been creating a further divide in opinions. To resolve this divide and let users understand both sides of the news, we have created Nexus News.
Solution:
NexusNews is a news app which integrates different news from all over the internet. Our project aims at making users aware of the biasedness of the they read. This provides users clarity of information on the news while also encouraging them to explore different perspectives.
Bias Prediction-
We categorise the news article into 3 perspectives for politics namely left sided, Centered and right sided.
We have trained our model such that it takes in news article as input and returns the output as percentage of bias.
Sentimental Analysis-
We take our project to the next step by implementing a new feature for non political content namely Sentimental Analysis. Through this, we have implemented the emotional leaning of the headlines of the news as a percentage i.e whether the given news is perceived as positive news or negative news or neutral news.
Pay per view-
Adding on to this stack, we introduce our web3 integrated pay per view feature. Through this feature, we aim at providing news journalist and writers an opportunity to increase the monetary value of their articles by blogging exclusive content sheilded behind our pay per view encryption feature .
This will help them make a better stand for themselves and achieve a better monetary value for their articles compared to traditional news websites but also make viewers happy by declutterring biased and untrustworthy news as the users will only continue to pay as long as they are provided with key valuable insights .
Tech Stack & Implementation:
We have created the News webapp in MERN (MongoDB, Express,React,Nodejs) stack.
To import the news , we have used the free version of newsapi.org.
for ML we have used Flask API to create backend and deployment for our ML models that include sentiment analysis and bias prediction as the key components of our project.
We used Figma to develop the UI/UX design and guide our process during frontend devolopment.
We used Web3 solana smart contracts to implement the pay-per-view feature.
The project was built close to industry standards with clean and professional code.
Github repo Link:
https://github.com/Pritish2005/NexusNews
Devfolio Link:
https://devfolio.co/projects/the-nexus-news-app-an-aipowered-news-platform-3ebf
Presentation Link:
Demo Video Link:
https://drive.google.com/file/d/1oIEB4FxBSxz29yAUYCUdiEXIJqtDIubI/view?usp=sharing
Overall Experience:
We got this development problem which we had to solve by integrating a few APIs and making it flow smoothly.
On one hand we needed to keep the information flowing in real-time but on the other hand we had to make sure performance was not compromised.
After checking and refining the bias prediction and sentiment analysis models, we made sure that they were valid and efficient.