Vice President Leni Robredo's supporters have filled parks and stadiums with tens of thousands of supporters to show that while she's behind in opinion polls, she can bring warm bodies into the streets. Whether they're pushing to install or oust their leader, the voice gets louder as the crowd gets bigger.
On the ground, news outlets rely on policemen tasked to guard the event as well the organizers to estimate the number of people at a rally -- usually based on the size of the venue (which is exact) and the density of the crowd (which can be subjective).
Often, the estimates from authorities and organizers vary. However, there are online tools that harness the power of Google Maps to estimate the size of a crowd. While science-based, there are nuances on the ground that a satellite snapshot can't catch -- like people crowding under shades, climbing onto trees and bleachers, and those who are positioned in side streets that are a few steps from the main venue.
Take for example the record-setting six million people that went to Pope Francis outdoor mass in Manila on Jan. 18, 2015. While the main venue was Quirino Grandstand in Luneta, people spilled over into Roxas Bolevard and sidestreets that fan out to Taft Avenue.
The day before, Francis celebrated mass for an estimated 200,000 people in Tacloban City at a venue that's much smaller than Quirino Grandstand -- a parking lot right behind the Daniel Z. Romualdez airport. What made the crowd so dense? Wearing yellow ponchs because of a passing typhoon, they were sealed behind metal railings thus preventing tens of thousands from moving around.
In recent years, one of the biggest rallies at Luneta was outgoing President Rodrigo Duterte's Miting de Avance or final campaign rally in 2016, drawing some 350,000 people. Veteran policemen and journalists covering rallies in the area know that a crowd is big if it breaches the carabao statue, meaning it has swelled from the bay to Roxas Boulevard.
On Aug. 26, 2013, police estimated some 60,000 people gathered at Quirino Grandstand to protest then President Benigno Aquino's handling of "pork barrel" or discretionary funds for congressmen.
Why big crowds matter
With Metro Manila and some urban centers under the loosest set of COVID-19 restrictions, presidential candidates have begun organizing large rallies after their day-long motorcades. There are two months left in the campaign before the May 9 vote.
In a September interview with reportr, University of the Philippines political science professor Ma. Ela Atienza said these classic shows of force help politicians bring voters to their fold
“Candidates have to convince people or voters that they are strong candidates and they have the capacity to win,” she said.
The people on and off stage at political rallies symbolize who they represent and why they want to run in the first place, said political analyst Cleve Arguelles.
“If you're branding yourself as a pro-poor politician, then you're likely to see poor people being mobilized to attend the event. If you're selling yourself as a regional candidate representing the interests of a region, then you'll see people from that region as well,” Arguelles, who teaches political science at the De La Salle University, told reportr.
“It isn't really a show of force but a way to tell one's campaign story,” he added.
Cavite Congressman Crispin "Boying" Remulla, reacting to the turnout of the Robredo rally in his province, said some of the attendees were paid and bused in. His family supports Robredo's rival, Bongbong Marcos.
Why giving crowd estimates is tough
There's a danger for crowd estimates to become inflated or deflated due to politics or simply the failure of the one giving it to properly approximate the size, said Stephen Doig, a data journalist and journalism professor at Arizona State University
They key is to cross-check the estimates from the ground with satellite imagery and math, he wrote on The Conversation.
"It’s actually a difficult thing to do. Once you get more than a few hundred people, it just looks very big. All you can do is pull big numbers out of the air," he said.
Doig said estimates should be based on the size of the area in square feet or square meters then give a "reasonable estimate" of the crowd density, or how much space each person is taking up. This is where the estimate could go far from the actual size.
Modern crowd-size estimation techniques are typically based on the Jacobs Method, named after a journalism professor at the University of California, Berkeley, who was watching Vietnam War protesters outside his office window, and noticed they were standing on a paved pattern of repeating squares.
He counted the students in a few squares, and calculated the average number of students per square, or crowd density. Then he simply multiplied the number of squares by the density to estimate the size of the crowd.
From his observations, he found that in a light crowd, each person takes up about 10 square feet (0.93 square meters), whereas in a denser crowd each person occupies less than half this space. In the most densely packed crowds, each person occupies just 2.5 square feet (0.23 square meters) – referred to by researchers as “mosh-pit density”.
This is considered an upper limit to crowd density, because it is not physically possible for a person to occupy less space. Hence, any crowd estimate that assumes a density higher than that of a mosh pit can be safely discarded.
This basic principle is used by some online tools to estimate and fact-check the number of people standing in a given area. Instead of counting squares, the total area is multiplied by the density to calculate the crowd size estimate.
Although these tools give a decent rough estimate of the total crowd size, they assume a uniform distribution of a crowd across an area, which is not realistic. This method also fails to take into account the space taken up by street furniture, cars, trees, or other spaces not occupied by people.
AI algorithms can count people by recognizing and counting the distinctive shape of humans, or even just their heads in denser crowds. Statistical methods can also be used to detect the independent motion of the people in the crowd. Or, if the crowd is too packed to count individuals, groups of people can be tracked.
It’s harder to estimate the size of a mobile crowd than a static one. The crowd density of a political march can vary significantly as people join and leave at various points along the route, and banners or placards can make people effectively invisible to crowd-detection algorithms.
Some researchers suggest using on-ground inspection points where people are counted. The best estimates are likely to involve multiple complementary methods, such as direct counting, aerial and map-based imagery, and public transport data.
Of course, knowing the size of a crowd is about more than just earning bragging rights for politicians. It is a crucial part of crowd management and safety monitoring at large events such as sports fixtures and music concerts.
-- with reports from The Conversation via Reuters Connect