Facebook's mission statement is simple - Give people the power to build community and bring the world closer together. But with the increasing threat of false information, along with the rise of extremist groups that promote toxic ideologies, we all have become more vulnerable to echo chambers which may provide misleading information and narrow, one-sided perspectives.

This is why we built Bubble - A factual, compassionate, and bipartisan news chatbot.

What it does

This chatbot aims to allow users to naturally communicate and receive emotional support along with factual bipartisan information on their topic of concern. It provides the emotional, intellectual, and political intelligence required to discuss complex topics that have widely differing perspectives across the political spectrum in a representative and empathetic manner.

How we built it

This chatbot was built utilizing using the following open-source packages and APIs:

  • DeepAi's Sentiment Analysis API developed by Stanford NLP to identify the emotional tone of messages.
  • keyword-extractor by Michael DeLorenzo, a keyword extractor API which would be used to identify the main topics within a message.
  • puppeteer, a Node.js library for automating, testing, and scraping web pages on top of the Chrome DevTools Protocol.
  • Cheerio, an API for traversing webpages by parsing the markup results from the puppeteer request

Challenges we ran into

The main challenge was correctly implementing the sentiment analysis and text extraction Natural Language Processing libraries to provide meaningful context to the users' messages. Along with finding the balance of providing the user with enough information to understand an array of perspectives, but not so much information that it becomes overwhelming.

Accomplishments that we're proud of

We built a working prototype that accomplished all the main objectives we set out to achieve.

What we learned

  • News, perspective, emotional intelligence, and information are very powerful tools that can either empower communities and individuals by promoting healthy conversation and growth, or prey on our biases and vulnerabilities to promote hate and fear.
  • Natural Language Processing still has a long way to go with regards to understand emotion and contextual speech.

What's next for Bubble

Continue developing Bubble into a natural, informative, and reliable news source which encourages open communication and promotes compassion

Share this project: