Maithilee Sagare

PuneBusLive – Smart Commute Assistant

Daily bus commuters often face uncertainty while traveling. Existing transport applications provide route information but do not answer practical questions such as which bus to take, whether the bus is delayed, whether seats are available, or if waiting for the next bus would be a better choice. This results in missed buses, overcrowding, longer travel times, and an unpredictable commuting experience. The challenge was to create a solution that helps commuters make informed travel decisions instead of relying on guesswork. Process As a resident of Pune, I have personally observed the challenges faced by PMPML commuters, including unpredictable bus arrivals, overcrowding, and limited real-time information. Since I was familiar with the city's public transport system, I selected Pune and PMPML as the implementation environment for the project. I researched common commuter pain points and identified that users needed decision support rather than just route information. Based on these findings, I designed PuneBusLive as an AI-powered commute assistant. The system evaluates routes using Speed, Comfort, and Reliability metrics to recommend the best travel option. I implemented route discovery for direct and transfer journeys, developed simulated live data models for ETA prediction, traffic conditions, delays, and occupan Results PuneBusLive successfully converts public transport information into actionable travel recommendations. The platform supports 24+ PMPML routes and 35+ bus stops while generating route rankings based on ETA, comfort, and reliability. The recommendation engine produces commute confidence scores and seat availability predictions that help commuters make informed decisions. The solution demonstrates how AI can improve the public transport experience by reducing uncertainty and providing clearer travel guidance. Reflection Through building PuneBusLive, I learned how to solve real-world transportation challenges by combining user needs with technology. I gained experience in route recommendation logic, ETA prediction, seat availability estimation, and responsive web design. The project showed me how traffic, occupancy levels, route reliability, and seat availability influence commuter decisions. Given more time, I would integrate live GPS feeds, real passenger occupancy data, and machine learning models to improve ETA, delay, and seat availability predictions, making the platform more accurate and scalable.