

This recent NY Times article about crowd estimates for Hong Kong's annual pro-democracy protest is a good follow-up to my post about the number of people who, allegedly, showed up to last month's NBA Championship parade here in Toronto.
For years, Hong Kong has been seeing divergent estimates for its annual protest. Organizers typically overstate. And the police typically understate. This year, organizers claimed 550,000 people in attendance, whereas the police claimed only 190,000.
The difference this year is that a local tech company has started using AI software (loaded up onto iPads) to help supplement the standard practice manual counts. This year they concluded -- perhaps more definitively -- that 265,000 people protested in the streets of Hong Kong.
Image: NY Times

Today was a historic day for Toronto, for Canada, and for the game of basketball in this country. The Toronto Raptors are world champions for the first time since their founding in 1995. Soak it in. Here is a photo that I took of the parade coming through the Financial District at around 2:30pm:

Some of the estimates going around are that 1 to 2 million people attended today's championship parade. But 2 million seems like a lot, even though today was frenetic (see above photo, again). I mean, that's 1/3 of the population of the Greater Toronto Area.
The fact that some of the "official" estimates also have a 1 million person spread tells me that, as of right now, we actually have no idea how many people were at today's parade.
So that got me thinking: How do people count crowds? And are we using drones to do it, yet? Subway and rail ridership for the day -- which surely spiked -- will give us some indication. But definitely not the full picture.
It turns out that the typical approach to counting crowds is known as Jacobs' Method. It was invented in the 1960s by a professor at UC, Berkeley, named Herbert Jacobs. He came up with the method while trying to count the number of students protesting the Vietnam War.
The concept is simple: It's area x density. And permutations of his method usually use this same principle. What you do is take the area filled with people, break it up into a smaller grid, and then come up with a population density estimate for each square.
He had some rules of thumb for that. A light crowd was about 1 person per 10 square feet. And a dense crowd (such as a mosh pit or an NBA championship parade in Toronto) was about 1 person per 2.5 square feet.
Using this method and aerial photos of today's parade, I would imagine that we could eventually get to a more precise estimate than 1 to 2 million people. But surely somebody has figured out how to program a drone (or other UAV) and do this even more accurately.
Crowd data is valuable information, particularly for political rallies and protests (I would imagine). If you know of a company doing this, please leave it in the comment section below. And if it doesn't yet exist, well then, now you have a new business idea.