5 Q’s for Florent Peyre, Co-Founder of Placemeter
The Center for Data Innovation spoke with Florent Peyre, co-founder and chief operating officer of Placemeter, a startup in New York City that uses live video feeds and computer vision algorithms to measure activity in public spaces in real time. Peyre discussed the opportunities to use automated video analytics to better understand cities.
Daniel Castro: What is the main problem Placemeter is trying to solve?
Florent Peyre: Technology is rapidly changing how people interact with the physical world. We count calories burned with a wristband, order a car service with a button, and even track our friends’ whereabouts with our smart phones. At the same time, as urban living becomes more desirable, citizens and cities are increasingly interested in getting the same sort of instantaneous information from the space around them: how many cars speed in their neighborhood, how crowded the park is on Sunday morning, how many pedestrians walk by their family bakery every week.
Unfortunately, up until now, the data that drive these insights have been unreliable and difficult to obtain.
Placemeter believes that a city will operate better with more data. By using a collaborative effort between citizens, cities, and businesses, Placemeter generates and distributes data that’re essential to realizing the vision of smarter, more accountable cities. Knowing how people move about and use a city will fuel innovation to help businesses, residents, and governments better understand the world around them and, conversely, help cities be more cognizant of these institutions.
Castro: How are your customers using the metrics you are able to extract from live video feeds?
Peyre: We have a customer base with a diverse range of interests in our data. Retailers and local businesses utilize Placemeter’s data to understand how busy their neighborhood is and to know how much foot traffic is walking by their location. Urban planners and local governments use our data to precisely measure the needs of a city, answering such questions as: Is a park under or over used? Is a sidewalk too narrow or congested? Is this street optimized for pedestrian safety? Economic development agencies also use our data to attract retailers to emerging neighborhoods and to measure the impact of their development and zoning efforts on foot and vehicle traffic.
Individuals and grassroot groups also want the data. Civic hackers (urban data geeks), neighborhood associations, or just normal citizens want to understand how their cities work. For example, someone contacted us to get foot traffic data about their newly popular and hip neighborhood: they see a lot more visitors, but the city has not installed more trash cans, so they also see a lot more trash. Placemeter’s data gives concrete proof of their arguments to the city. Pedestrian safety activist groups also want to use our data or install sensors to measure the speed of cars and leverage that information with their local community boards to get speed bumps installed and traffic controls implemented at the right locations.
That’s not it, though. We’re discovering many more uses for our data every day. We’re excited to see the imaginative things other people and groups will do with Placemeter as it grows.
Castro: Automatically counting pedestrians and vehicles in video feeds seems like just the beginning. What other applications do you see for applying algorithms to video feeds?
Peyre: Placemeter’s goal is to understand how people use cities and we’ll continue developing the best algorithms to answer that question. The goal is for the technology to be able to teach itself—much like a baby recognizes that a mug and a glass, although not exactly the same shape, both serve to hold liquids. We are wading into unknown territory of what important patterns might be recognized and how cities can use them to improve the everyday lives of its citizens. Algorithms like ours will become ever more important as cities grow, so it’s important that they’re developed alongside a comprehensive understanding of their privacy implications. At Placemeter we design our algorithms with privacy safeguards built-in, like making sure we don’t use face recognition technology. Placemeter has instituted a variety of common sense principles—such as not recording over 99% of the video we ingest and conducting regular reviews of our policies and practices. Technologies like ours will only flourish if they respect the basic level of privacy that people expect in their everyday lives.
Castro: This field seems to be developing rapidly. Where do you expect to see the most innovation: around the sensors for capturing images or in the algorithms for processing them?
Peyre: Both! Algorithms will indefinitely become smarter and learn to use more data to improve themselves, whereas the sensor hardware can be better tailored for the algorithm’s specific needs. The field has so much potential, it’s exciting to be in the middle of it.
Castro: How will communities change once we are able to quantify more about city life?
Peyre: There will be a public data layer which does not exist today. In the future, once Placemeter’s data and other’s are more widely available throughout the city, you can imagine a world where you know exactly when to go grocery shopping and avoid the line, when it’s best to go to the museum in a new city you are visiting, which cafe to go to to get your coffee faster, which bar is busy and which one is more quiet, and to find an available court for basketball with your friends—the use cases go on and on. Everybody will have the power of a city’s pulse in the palm of their hand.