Skip to content

Latest commit

 

History

History
76 lines (62 loc) · 4.79 KB

README.md

File metadata and controls

76 lines (62 loc) · 4.79 KB
Raw

Failed-In

Flaunt Your Failures!

✨Succeeded in winning the Most Funny Hack in HackHarvard 2021✨

pic

Table of Contents

Inspiration

On social media, most of the things that come up are success stories. We've seen a lot of our friends complain that there are platforms where people keep bragging about what they've been achieving in life, but not a single one showing their failures.

We realized that there's a need for a platform where people can share their failure episodes for open and free discussion. So we have now decided to take matters in our own hands and are creating Failed-In to break the taboo around failures! On Failed-in, you realize - "You're NOT alone!"

What it does

  • It is a no-judgment platform to learn to celebrate failure tales.
  • Enabled User to add failure episodes (anonymously/non-anonymously), allowing others to react and comment.
  • Each episode on the platform has #tags associated with it, which helps filter out the episodes easily. A user's recommendation is based on the #tags with which they usually interact
  • Implemented sentiment analysis to predict the sentiment score of a user from the episodes and comments posted.
  • We have a motivational bot to lighten the user's mood.
  • Allowed the users to report the episodes and comments for
    • NSFW images (integrated ML check to detect nudity)
    • Abusive language (integrated ML check to classify texts)
    • Spam (Checking the previous activity and finding similarities)
    • Flaunting success (Manual checks)

How we built it

  • We used Node for building REST API and MongoDb as database.
  • For the client side we used flutter.
  • Also we used tensorflow library and its built in models for NSFW, abusive text checks and sentiment analysis.

Challenges we ran into

  • It was the first time we tried using Flutter-beta instead of React with MongoDB and node. It took a little longer than usual to integrate the server-side with the client-side.
  • Finding the versions of tensorflow and other libraries which could integrate with the remaining code.

What's next for Failed-in

  • Improve the model of sentiment analysis to get more accurate results so we can understand the users and recommend them famous failure to success stories using web scraping.
  • Create separate discussion rooms for each #tag, facilitating users to communicate and discuss their failures.
  • Also provide the option to follow/unfollow a user.

Contributors

Images








Built At HACK HARVARD