Inspiration
Our collective interest in reading news articles and implementing Natural Language Processing methods were a great inspiration for the project. We were also motivated by the problem that most people in try to avoid reading negative articles. We came with a creative solution to solve all these problems
What it does
Don't have the time to read the full article? Trying to avoid reading negative news articles? This project is made just for you! After pasting the link of the the article, it will create a summary and classify if it is a positive or negative article. It will also give the users a similar NYT article to the article pasted. This app will also give a the most common tags found in the article that are relevant
How we built it
We used python modules from BART created by Hugging Face. We then worked on classifying the most popular tags of the article by getting a score that was implemented using a transformer by hugging face. We then worked on getting the NYTimes API to get a similar result which included some words of the headline of the original article if a reader wants to know more about the topic discussed in the article. At the end we used Streamlit.io to create a creative and responsive web app and implemented our project in there.
Challenges we ran into
We tried to implement this project on cloud by using either GCP or Streamlit cloud. Unfortunately, we were unable to get those implemented in the given amount of time. We also ran into issue of sorting the tags and presenting the user the most relevant tags of an article.
Accomplishments that we're proud of
Completing the project was a great achievement for us! We are also proud of the accomplishment of using NLP models in a hackathon.
What we learned
How to use BART by Hugging Face (NLP library)
What's next for Articles NLP
Checking the political bias
Log in or sign up for Devpost to join the conversation.