Meeting Time:
Thursday 7:00-8:15pm
Project Description
The Predictive Crowd Flow/Crush Analytics system will help event organizers, and fire marshals to be proactive in preventing crowd crushing incidents at large scale events. Current methods are not effective at monitoring and predicting dangerous crowd dynamics in real-time. Past incidents such as Astro world and the Who concert have shown there are precursor signatures to dangerous situations, such as bottlenecks and wave-like crushing movements. The proposed system will monitor the large scale event and automate alerts when risk metrics exceed a threshold, thus reducing the criticality of human observations. Design a tool that users can upload event floor plans and input variables such as estimated attendance, entry/exit points, key locations, the system could simulate and identify potential bottlenecks prior to the event, allowing for adjustments for safer layouts.
Relevant Skills: Video Analytics, Machine Learning, Simulation/Spatial Analysis, Backend Development, Frontend Development, Crowd Science, Event Planning, Data Science