Erin Morrow's MassMotion Simulates Crowds
Over the past decade, Erin Morrow has become an expert on the movements of the masses. He’s a transportation planning consultant and focuses on helping architects understand how crowds will flow through the airports, hotels, and skyscrapers they dream up. Morrow predicts how many people will line up at Starbucks during commuter hours and determines where to place an escalator to minimize bottlenecks.
Morrow ’s analytical weapon is a software package he created called MassMotion. It takes 3D computer models of buildings and fills them with “agents,” virtual stick figures that represent people. These avatars aren’t mindless and in fact have specific personalities. Some meander. Others are busybodies in a rush. The agents will linger in front of signs and puzzle over directions, and they’ll also grab smokes together outside of exits. “They all have certain things they care about,” says Morrow.
While studying at the University of Waterloo, Morrow fancied himself a budding architect. Yet he dropped out and went to work at the Toronto office of Arup, a London-based design and engineering firm that specializes in large projects such as the Sydney Opera House and the Bird’s Nest stadium in Beijing. Morrow, now 34, tapped into an innate understanding of geometry to begin advising colleagues on the layout of buildings. “I don’t know how to describe it,” he says. “I can just move things around in my head.”
Morrow began working with 3D software and taught himself to code. When Arup started work on New York’s Fulton Street Transit Center eight years ago, he decided to write a software application to add a layer of intelligence to the crowd animations typically done for this type of project. MassMotion was born, and Arup has since used it on projects including the JetBlue terminal in New York and Sydney’s metro system. In June, Arup’s software arm, Oasys, started licensing MassMotion for $34,000 a pop.
To create an accurate model, Morrow gathers data including how many people are expected to pass through the structure, the arrival and departure times for planes and trains, and the number of stores and reception desks. MassMotion assigns individual tasks to tens or even hundreds of thousands of agents—check in for a flight, for instance, or buy a cup of coffee—and sets them loose. “We give them a range of walking speeds, body sizes, and preferences for, say, avoiding escalators,” Morrow says. One simulation for Union Station in Toronto predicted that a particular concourse would clog up as people got off their trains. “There were a few exit passages, but they led to a downstream choke point where people would end up waiting for about seven minutes to exit,” Morrow says. The architects added relief corridors and alternative routes. “It’s always interesting to see what the agents will end up doing,” Morrow says. “Their behavior will highlight issues that you would not have otherwise considered.”
The simulations aren’t just for architects: Building managers have used MassMotion to predict how evacuations will unfold, and the forecasts are generally within a minute or so of actual times. Police use the software to predict how crowds might disperse from venues. And concert promoters are trying to gauge the movements of showgoers at large festivals—which suggests Morrow may need to develop algorithms for drunken avatars.
During vacations, Morrow finds himself viewing cities in different ways as a result of this work. He was at first confounded by the chaotic roads of Italy, where traffic signs and rules are routinely disobeyed. He realized drivers pay close attention to cars directly in front of them, even while ignoring everything else. “It looks like madness, but it works as long as everyone does it,” Morrow says. “I like trying to figure out what makes these systems tick.”