FloodOps SG

FloodOpsSG is a web app that centralises flash flood information across Singapore into one cohesive dashboard, reducing reconciliation time first responders experience between the flood occuring and the responders being deployed. Additionally, a Deep Q-Network agent was trained using simulated sensor data modelled on live Singapore feeds — rainfall, water level, accumulated rainfall trends, and official government signals — and gets rewarded for escalating early and penalised hard for missing a flood event. The model weights are then exported to JSON to be integrated into our web app — no Python backend needed at inference time.