How many participated in candlelight protests? Statistical Counting of Dynamic Crowd

Abstract

The recent controversy on the size of crowd at candlelight protests raises a serious question on how to estimate the crowd size. There was a significant discrepancy between police and protest organizers on the size of crowds at each protest. While police often reports the crowd at its peak, organizers want to count anyone who participated in the event from its start to finish. We propose a new counting method using an inverse probability weight estimator. Assuming that the arrival and departure times of crowd can be estimated using additional survey, we estimate the probability of a subject being in the events at a given time. We also propose a bootstrap procedure to obtain the variance of the estimator. We demonstrate the performance of the proposed method with simulation studies and the data from Koreas March for Science, a global event across the world on Earth day April 22, 2017.