EV ChargeSmart is a web platform that reduces waiting and congestion at EV charging stations in India.

The problem: Public chargers are scarce (about 12,000 now versus 400,000+ needed by 2030, per NITI Aayog). Drivers often wait more than 20 minutes at peak hours, yet average charger utilization is only around 55%. There's no single platform that combines live charger monitoring, congestion prediction and advance booking.

What it does

Live station map: shows charger availability and status across the network, fed by IoT sensor data.
Congestion prediction: Random Forest, LSTM and GRU models forecast how busy a station will be over the next 30 minutes to 3 hours.
Slot booking: you reserve a time and charger type in advance, with a reminder before your slot.
Smart redirection: if your chosen station is about to get crowded, it suggests a better one nearby.
Admin dashboard: gives operators fault alerts, utilization analytics, peak-hour demand and charger management.
Extras: a charging estimator for time, cost and range, favorites, check-in, and a personal impact summary.

How it's built: a four-layer design.

IoT data layer: a simulated MQTT sensor feed stored in MongoDB and Redis.
Intelligence layer: Python and Flask serving the ML models.
Application layer: Node.js and Express, with JWT login and admin/user roles.
Presentation layer: the web portal and admin dashboard.

Targets and results shown on the site: a 40% reduction in wait times, 80% charger utilization, RF accuracy of 88.4%, and LSTM and GRU errors of 2.81 and 2.94 kWh.

The website you have is a front-end demo of this idea. It uses sample data, and the ML models and backend aren't running behind it. If you want, I can help you explain that clearly in a viva or report.