Think back to your elementary school days. Have you ever had a moment where a group of kids in your class verbally harassed you? Been accused for something you didn’t do? Or even get straight-up bullied? But because of the torrent of social pressures, you couldn’t bring yourself to tell someone. Or even worse, you had told your teacher, but for a variety of reasons, they either won’t believe you, or they won’t take action. As ridiculous and infuriating as these stories are, these exact situations occur everyday in schools across America. In fact, 1 in 5 kids reported instances of being bullied.

So how could we protect these kids from this fate, and prevent long-lasting emotional scars that could affect them for the rest of their life? And how can we prevent offenders from continuing down a slippery slope, and have them take responsibility for their actions? Especially if teachers and administrators lack the time and resources to do so effectively?

What it does

SayfeZone aims to provide an affordable and highly scalable monitoring system that can be spread across the school. Through cheap Android devices or other camera-enabled IoT devices, a network of devices can beam video to our database, which will be run through an NLP processor and sentiment analysis. Highly negative instances trigger a notification, and gets marked on an administration panel as a moment of importance.

Administration can then evaluate at their leisure, perhaps daily or even weekly, taking appropriate action based off what they've seen. In worse-case scenarios, the footage provides legal documentation, allowing justice to proceed. With irrefutable proof at who could be at blame, this system is meant to foster a safer and honest environment. This system could be further extended to identify children facing social difficulties, and with the same technology, extended to catch school shooters earlier by tracking weapons.

Due to security concerns, data not flagged as concerning will be deleted from the system in a timely manner.

How we Built it

See above for an idea of our (planned) system flow! Technologies used include: React Native for recording video/audio Google Cloud for speech-to-text, NLP, and sentiment analysis jQuery/javascript front-end administration panel w/ C3JS for graphing S3 databases to store video ...and of course stdlib, which allowed us to seamlessly integrate our API calls and data management with our two front-end points, without needing a dedicated back.

Challenges we Ran into

  • Figuring out how to record and encode video in base64 while chunking for potential streaming made us run into many incompatible API issues and made us pivot front-end technologies in order to provide a high-quality stream

  • Handling large amounts of data as well as processing presented several issues, such as ensuring time-lag does not negatively impact user-experience, while also adhering to database storage difficulties and server timeouts

  • Sequencing issues due to varying packet speeds as well as sequential vs asynchronous calls

  • Despite our well-thought out implementation plan, many parts could not easily be coordinated and resulted in blockers; for example, the core data was not available for much of the front-end testing and streaming

  • Learning new technologies and techniques to over come design challenges, such as smoothly displaying video chunks as a stream

Our Accomplishments

Through perseverance, we believe our team has managed to face these challengers head on. Despite the difficulties, with the help of some talented advisers, we believe we've developed a solid prototype. By doing so, our team feels that each of us were able to learn something new. We dived into various niches that we believe we never would have tried otherwise: from the intricacies and benefits of AWS, to experimenting with new mobile APIs, and overall developing new skills that we have no doubt will help for future ventures.

What's next for SayfeZone

We thoroughly believe in our core mission in making our education system safer. With some our members having having previously experienced such traumas, either directly or through family members, we want to start taking the next steps and learn about potential challenges with future integration. As a base product, SayfeZone has many applications and extensions, and our team is very much interested in seeking further opportunities to expand our vision into the real world.

Share this project: