-
-
The UI
-
-
The page you are greeted with
Inspiration
Our main inspiration was realize how cumbersome the process of doing background research on a story can be sometimes. Especially when a news story has been covered for a long period time, making it difficult to see how the story has developed over time.
What it does
The user enters a news article that they are reading. Our app then finds out what the article is about, and looks for other relevant articles from the past. It then gives you almost a sort of reading list to catch up on the story so you are in the loop and have background information.
How we built it
Our back end is running a Python flask server. We noticed that almost all news article webpages have the headline in the Title tag for the webpage. So to retrieve the headline we just get the HTML code, parse it, and get the Title tag's value. Then, we pass in the headline into the template for a prompt that GPT-4 uses to guess what topic the article could be about. Finally, we use the topic that GPT-4 thinks the article is about to get past articles that could be relevant using the NewsAPI.
Challenges we ran into
Currently, we can only search articles at most a month in the past. Searching back farther requires a paid subscription to a higher tier for the NewsAPI, and we don't have the financial means to pay for it. So our app can only accurately compile a list of relevant news articles for very recent stories. Another problem is that sometimes the headline of an article misleads GPT-4 and our prompt gives us a topic that isn't actually relevant to the article. Because of this, sometimes the list of articles that you get can be very different from the original article and our app isn't so helpful.
Accomplishments that we're proud of
We all learned a lot, since three of the four memebers of the team only know the python we were taught in CSE 115, however, a lot of research and learning was put in to figure out how we can apply our knowledge to help build the backend of the project using various python libraries. Overall, we are proud of trying to expand our knowledge and applying it to this project. We had a lot of fun learning more!
What we learned
We learned how to read the doucmentation for how to apply python libraries and use those libraries to accomplish our project's goal. We learned how to implement APIs and we also learned how to create our own! Which is something we didn't even think we could do at first.
What's next for Contextification
We want to expand on our app's functionality, by giving our app the capability to look back farther than a month to find articles. And we also want to improve the accuracy and relevancy of the article search results so that the user can find more useful background-information.
Log in or sign up for Devpost to join the conversation.