The COVID-19 global pandemic has a great negative influence on the event industry. Decreasing number of festivals poses threats to cultural and economical development. This research aims to close this research gap by developing a SARS-CoV-2 transmission risk analysis method for large events by modelling crowd interactions at different types of event spaces and quantifying the SARS-CoV-2 transmission risks in the process. A method is proposed to connect activity scheduling, pedestrian route choice and movement modeling, virus spread modeling, and infection risk identification to determine the transmission risks throughout the events by probability method.

A case study applies the proposed method in a pre-pandemic music festival and shows its capability of revealing the general infection risk and the relation of influence factors to the transmission scale, and it also identifies risk-prone areas in an event. Comparing the different scenarios at each activity space, the general trend is identified that the transmission of SARS-CoV-2 is limited when the facility is located outdoor, the queue distance is increased, the density is lowered, and the respiratory activities are calmer. Compared to the real-life experimental events, the simulated results tend to be underestimating the risks due to assumptions of the event scale, infrastructure, compliance to measures, heterogeneity in the emission rate, activity schedules, group behavior, number of infectious individuals, and such factors that directly influence the amount of virus transmitted in the event. Nevertheless, this research has shown when and where major risks can occur during an event. It also gives indications on crowd management measures and interventions that can help reduce the virus transmission scale.

The proposed method allows future exploration and comparison of the transmission scales of large events, without posing ethical controversy of exposing people in infection risks. Yet, it has its limitations of not considering group behavior at the event, which is commonly observed at large events and may potentially increase the transmission scale. Another major limitation is its heavy dependency on detailed virus transmission parameters and activity patterns. The former needs to be validated under different scenarios, and the latter needs to be identified from the data collected at the same type of event in future research.

This research explored two types of models, ARIMA and multiple linear regression, for forecasting tourist counts in 7 locations around Amsterdam red light district, for the prediction horizon of up to 30 minutes.