TeenGram: A Safe, AI-Driven Social Media Platform and Hybrid Recommendation Engine for Youths

Modern social media platforms facilitate vast information exchange and community building, but they also expose youths to significant online risks, including cyberbullying, harassment, and harmful graphic content. This paper proposes the architecture and methodology of TeenGram, a novel social media web application specifically designed to provide a secure, filtered digital environment for youths. By integrating a strict identity verification process via college administration bodies, TeenGram eliminates the risks associated with online anonymity. Furthermore, the platform utilizes a dual machine-learning (ML) model approach for real-time automated content moderation and feature extraction. To deliver high-performance data processing, the system architecture leverages Next.js, MongoDB, and Redis as an in-memory data store. Finally, to ensure high user engagement, TeenGram implements an advanced Hybrid Recommendation Engine that synergizes Collaborative Filtering, Demographic-Based Filtering, and Content-Based Filtering.