Event Horizon

• 

  initial search and GPT generated response with table to articles

• 

  Cards of articles that include image, headline, author, and source

Inspiration

From the advent of AI to the recent conflict between Israel and Palestine, there have been many good, bad, and definitely important headlines in the news. However, with all the attention-seeking headlines and political pundits floating around, educating oneself on these topics while skirting biases could be daunting. All too often, people will only see one side of the story. Our website aims to directly address this situation.

What it does

Event Horizon curates headlines related to a topic of interest from both sides of the political aisle and delivers them to the user. These headlines are taken from both recent and past times to give more complete context, and a GPT-generated summary of discourse around the topic is also provided to aid the user's understanding. One thing that makes our website unique is its ability to curate articles of the past as well. Many current events have a complex and nuanced history that lead to the current situation which many current headlines fail to address.

How we built it

Event Horizon utilizes the React framework, NewsAPI.orglink, and ChatGPT. The user's query is sent to both NewsAPI and ChatGPT. The first returns a list of new articles related to the query, half from CNN and the other half from Fox, the latter is asked to "write a summary of discourse surrounding (the query) over the years." The responses from both APIs are then displayed.

Challenges we ran into

Most of us were unfamiliar with React.js coming into the hackathon. As a result, we made a lot of mistakes in the way we organized our project, and rendering our website upon data updates turned into a nightmare. However, we regrouped and took the time to understand the logic behind React. After we understood what was going on, we realized what we did wrong, and took the time to fix our code.

Accomplishments that we're proud of

Despite facing a great setback on the first day, through an all-nighter, we were able to bounce back and create a website better than what we had originally planned. We are very proud of our ability to keep our morals high and preserver through our mistakes and learn from out experiences.

What we learned

Don't get too hung up on our mistakes, learn and move-on. We also learned not to rush into projects and how important it is to understand the foundations of what we were dealing with. It never hurts to read up on documentation, and even build a few practice projects with the tools we plan to use beforehand.

What's next for Event Horizon

Because NewsAPI charged money for searching for older articles, we only used recent articles for our demo. However, a more developed version of Event Horizon would naturally have articles going decades back. This more developed version would also feature articles from a much larger variety of sources, and allow users to rate where an article falls on the political spectrum to further ensure equity. Finally, we'd also explore prompt engineering with ChatGPT to see what kind of prompts in relation to user inputs generates the most informative answer.

Built With

• chatgpt
• css
• gpt
• html
• javascript
• newsapi
• react.js