CERBERUS Emergency Response

Inspiration

220 active shooter events between 2000-2016 and 1,486 people were either killed or wounded
Statistics show that the majority of shooters will not stop until they are confronted by law enforcement
Average active shooter incident lasts 12 minutes or less, and the first calls to 911 do not typically come in for a number of minutes
- No application like it in the market

What it does

1) Allow an user to report active shooter and request medical assistance using an iOS app with the click of a button 2) Collect geolocation and sound* from users reporting active shooter to determine the shooting event and active shooter location *not implemented 3) Facilitates reporting of an active shooting event and decreases first responder's response time

How I built it

Built an iOS app with React Native to collect user's geolocation and store it in SQLite database
Used Folium, a Python library that helps users create several types of Leaflet maps, to visualize the location of the users who reported active shooter and ultimately determine shooter's approximate location
Deployed Flask App with Folium Map and Chat interface to Google Cloud App Engine for First Responders

Challenges I ran into

Pivoted Project first idea Rewrote code from Swift to react Not enough time to implement TensorFlow

Accomplishments that I'm proud of

How far we came and the friendship we built

What I learned

A whole lot of technical things Resilience Teamwork Persistence

What's next for cerberus-rex

1) Collect real-time sound data once user opens the app and pipe to datalake in Google Cloud 2) Use TensorFlow, Google Audioset and Youtube-8M datasets to determine user's closeness to shooter given sound wave data 3) Implement user authentication