When we were brainstorming ideas for our project one of our team members brought up the Florida COVID tracker. In which a civilian used government data to track the spread of COVID-19 in Florida. This led to a rabbit hole of tracking tweets that happen during natural disasters.


What it does

Tradgy provides a heatmap of disasters and catastrophes in the United States. This is based off of a list of keywords that would used in such events.

How we built it

We used a script to pull tweets that have certain keywords, fed it through a neural network to determine the worthiness of an event, then convert the tweet location data to a list format and create a heatmap of it on a website.

Challenges we ran into

Creating a neural network, front end formatting

Accomplishments that we're proud of

  • Creating a neural network that uses NLP to discern tweet sentiment
  • Getting the map on the page
  • background on the 'how it works' page

What we learned

  • Wix has too many restrictions
  • Twitter is a hot pile of garbage (and so is their API)
  • Heatmaps are simultaneously great and awful
  • I wish you could compile JS code without jumping through hoops

What's next for Tradgy

Live data website

Share this project: