NAKSHA: Networked Analytics for Knowledge and Spatial Hub for Academics

Indian colleges still record attendance largely by roll call or paper sheets, which wastes lecture time, allows proxy marking and leaves the data unused until month-end, when struggling students are already far behind. This paper presents NAKSHA, a web system that needs no sensors or dedicated devices. A classroom screen shows a QR code that is replaced about every 20 seconds and expires after 90 seconds; students scan it with their own phones, and each phone is bound to a single roll number on first use. Identity, classroom and time are all decided by a database function, never by the phone. Each check-in updates one attendance table that drives both a live 2.5D map of classroom occupancy, which we describe as a digital shadow rather than a digital twin, and a mentor dashboard that groups students into support-framed bands from the last 14 class days. The system runs entirely on free hosting tiers using React and Supabase with row-level security, stores no protected attributes, and imports students only with recorded consent. We describe the design, the threats it addresses and our functional evaluation, and outline an explainable XGBoost and SHAP model as the next stage.