Inspiration
The shooting that happened this Friday at midnight was terrifying. The VT alerts notified us a lot later than the social media. Twitter has become an important communication channel in times of emergency. The ubiquitousness of smartphones enables people to announce an emergency they’re observing in real-time.
What it does
We aim at programmatically monitoring Twitter, use our Disaster Detection model to warn people about possible threats.
How we built it
We trained a deep learning model based on BERT using a public Twitter disaster dataset. And we created a website that could perform disaster statistics analysis, disaster tracking on Map, and disaster tweets classification.
Challenges we ran into
Stream real-time Twitter data
Accomplishments that we're proud of
We have an 83% accuracy on the disaster detention.
Built With
- google-cloud
- tensorflow
Log in or sign up for Devpost to join the conversation.