Proximity

Inspiration

My friend was once faced with a scenario where there is a school shooter in the proximity of the school, but had no ideas that it was happening in real time since the classroom were isolated. The shooter then came to another classroom and majorly injured many other people. It could have been him. This would not have happened if there is a warning system against nearby danger.

What it does

Proximity uses your phone sensor data in order to determine if you are in a panic state or not. If enough people are in a panic state and it overcomes a threshold, a warning message will send to everyone in possession of the app near your vicinity.

How we built it

We use machine learning in order to match data with a panic state, in this case, screaming and running. We train our data so that it recognizes if you are running because of panic or you just doing normal activities like jogging. We then use a mapping system in tandem with the GPS in order to mark your area and your "zone". We then use our training data set to test out various scenarios.

What's next for Proximity

Currently we are using a centralized system since we are short on time. However this would be very unrealistic since we do not want people to be sending their data to a third party that can be hacked at any given time. What we would want to do then is to use technology similar to block chain and develop a decentralized, peer-to-peer system so that everyone is anonymous and harder to take advantage of. We also are going to improve on our model and databases. Currently we only intergrated data for running and just finished training data for screaming. We also plan to develop it on the iOS platform.