The hi-res future of anti-pollution tech

Street View cars map the way to cleaner air, block by block.

Robert Ito

In the middle of West Oakland is a community garden where locals go to tend and sow small but cherished plots of squash, tomatoes, sunflowers, tree collards, pole beans, and kale. Plump hens take vigorous dust baths in the chicken coop nearby; honeybees from the garden’s hives and butterflies fill the air. For residents it’s an urban paradise, with beets.

Right outside, however, semis and concrete mixers and city buses whiz by at a regular clip. Down the street there’s a salvage yard filled with husks of cars, and an enormous recycling center where mini-Everests of scrap metal reach to the sky. Take a whiff in the garden and you’ll get notes of lavender and sage; on the street, it’s eau de truck exhaust. But if you were to ask most scientists or neighborhood activists about the air quality in this stretch of West Oakland, they wouldn’t be able to differentiate from one block to the next. They’d just tell you it was all lousy.

West Oakland Farm Park

  • Black carbon:

    High

    .6 micrograms per m³

  • Nitric oxide:

    Low

    6 parts per billion

  • Nitrogen dioxide:

    High

    18 parts per billion

Erin Cary, director of advancement for City Slicker Farms, at West Oakland Farm Park, where kale, squash, and tomatoes grow just a few blocks away from the perpetual gridlock on Interstate 580.

Until now. Beginning in 2015, a pair of Google Street View cars, equipped with high tech “mobile labs” developed by San Francisco–based startup Aclima, crisscrossed the streets of West Oakland taking second-by-second samples of the area’s air. They tested for nitrogen dioxide and a type of pollution known as black carbon (bad for your heart and lungs, not to mention the planet), as well as nitric oxide. The cars hit every stretch of pavement, from tiny cul-de-sacs to truck-choked Peralta Street, multiple times, taking millions of measurements.

There are three stationary air pollution monitors for all of Oakland, which reveal the city’s air quality as a whole. But the Street View cars can tell you what the air is like at, say, the corner of Market Street and Grand Avenue—basically anywhere you can drive a Street View car. They can even tell you how the air varies from one end of a single block to the other for a truly hi-res view of the problem.The result: one of the largest and most granular data sets of urban air pollution ever assembled in the world.

“We visited each block on between 20 and 50 different days over the course of a year,” says Joshua Apte, an engineering professor at the University of Texas at Austin. In the process, they were able to identify patterns they wouldn’t have otherwise seen. “If pollution spikes for an instant, it may or may not be such a bad thing. But if pollution is consistently high, that’s something we really should care about."

Anatomy of a mobile air pollution detection lab

  • 1. Aclima’s mobile platform is built onto Google’s existing Street View cars and makes use of their elaborate GPS tracking and 360-degree cameras.
  • 2. Air is sampled through an inlet on top of the car and pumped into a pollution-monitoring system in the back.
  • 3. Sensors measure substances like black carbon, nitric oxide, and nitrogen dioxide. Data on pollution levels for that specific locale are stored on the Google Cloud Platform and made available via Google Maps and Google Earth.

Project Air View brings together experts of all backgrounds and stripes. The nonprofit Environmental Defense Fund, which had previously partnered with Google to map methane gas leaks under city streets, brought Apte from UT Austin and other researchers together with Google and Aclima. Aclima had already worked with Google, measuring air pollution in Denver in 2014, before coming to the San Francisco Bay Area. To date, Aclima and Google have collected nearly a billion data points in California alone.

Aclima’s detectors are powerful enough to discern the tiniest amounts of airborne pollutants but small enough to fit inside a Subaru Impreza, and later in the Hyundai Santa Fes used in the Oakland study. (In the past, similar mobile monitoring projects in places like Beijing and Helsinki required large trucks and vans, often navigated by the harried scientists themselves.) Google provided the Street View cars and drivers, the cloud infrastructure to store the data, and the platforms (like Google Maps and Google Earth) to get the word out about the findings.

“The goal is to make this data available to more groups,” says Karin Tuxen-Bettman, program manager for Google Earth Outreach. “To put it in the hands of more people and make the invisible visible.”

Karin Tuxen-Bettman, program manager for Google Earth Outreach, wants to make street-level pollution data as accessible as Google Street View.

The researchers also monitored conditions in neighboring East Oakland and San Francisco, but West Oakland has been a particular focus of concern. Local groups like the West Oakland Environmental Indicators Project had long complained about the area’s air quality. Compared to the rest of Alameda County, West Oakland had nearly twice the rate of asthma emergency room visits and higher rates of stroke and congestive heart failure among residents. Unlike other sections of California, however, where the topography magnifies the problem—think Los Angeles or the Central Valley—nearly all of West Oakland’s issues come from local emissions. The neighborhood is ringed by three busy freeways—the 580, 880, and 980. Trucks rush to and from the nearby Port of Oakland, while the region itself is a mix of residential and commercial buildings, where single-family homes butt up against scrapyards and industrial warehouses.

West Grand near Peralta

  • Black carbon:

    Severe

    1+ micrograms per m³

  • Nitric oxide:

    Severe

    24+ parts per billion

  • Nitrogen dioxide:

    Severe

    24+ parts per billion

By combining location data with pollution sensors, researchers can detect exactly where pollution levels begin to spike as vehicles on city streets accelerate to merge onto the interstate.

As the data began to flow in, the researchers made a startling discovery: The pollution varied not just among cities or neighborhoods, but noticeably and dramatically from block to block. They also discovered specific points where the pollution was particularly high. Two of these “hot spots” are within sniffing distance of West Oakland’s Farm Park, home of all those chickens and bees and urban farmers; another is right next to the Willie Keyes Community Recreation Center, named after a beloved neighborhood activist who fought against, among other things, the region’s unceasing truck traffic. And the pollution didn’t just vary from block to block. Even on a single street, concentrations of pollutants could be five to eight times higher on one end of the block than the other. “In our scientific hearts we thought that might be the case, that pollution is hyperlocal, but we were still really excited to see the hot spots,” says Melissa Lunden, chief scientist at Aclima. "And then it was a matter of, OK, why are they there?"

From these findings, the EDF created interactive maps detailing the various pollutant levels and hot spots. Zoom in on a stretch of Peralta and you can see the point where cars speed up to get onto the Bay Bridge (thus creating more pollution) or where a steady stream of trucks head to and from the Port of Oakland. Colored dots that range from yellow (least bad) to dark red (highest concentrations) reveal each area’s pollution levels. “In this world where everybody’s got more and more data, the key is to be able to analyze it and move from data to information to knowledge to action,” says Millie Chu Baird, a senior director at the EDF.

rail tracks flowers

West Oakland residents had nearly twice the rate of asthma-related emergency room visits as residents of the rest of the county.

News of the study has already reached the community garden. Sitting on a bench amid plots of sunflowers and strawberries, Erin Cary, the farm’s director of advancement, describes taking the news—and maps—to her colleagues. On one map, the area by the salvage yard and recycling center (and, unfortunately, the farm’s playground) is peppered with dark red dots; as you move farther away, the dots cool to less-distressing oranges and yellows. “We’ve known that there are air quality problems, but we hadn’t seen it on such a microscopic level,” she says. “The air quality in this community garden is different than down by the playground, but you wouldn’t have any sense of that without those sorts of mapping tools.”

A block up the street at Willie Keyes, recreation leader Meredith Evans-Moore is supervising about two dozen kids, many of them playing basketball in the center’s large gym. She hasn’t seen the maps but isn’t surprised to learn that her rec center is right next to a hot spot. “We’ve got a lot of respiratory issues here,” she says. “More and more kids have asthma, bronchitis.”

Willie Keyes Recreation Center

  • Black carbon:

    Medium

    .05 micrograms per m³

  • Nitric oxide:

    Low

    6 parts per billion

  • Nitrogen dioxide:

    Medium

    12 parts per billion

This rec center in West Oakland is located just a couple of blocks from the busy Interstate 580 freeway. Meredith Evans-Moore says that her students who play here seem to have an increasing number of respiratory issues.

Project Air View is an ongoing endeavor, with similar trips taking place throughout San Francisco, Los Angeles, and the Central Valley. With each new outing, technological advances are exponentially increasing what people know—and what they can do with what they know. As next-generation air sensors become smaller and cheaper, researchers could place them on everything from city buses to taxis. Intelligent systems will be able to even more accurately forecast air pollution in real time based on a multitude of factors, from traffic patterns to that day’s weather to pollution-emitting businesses. “I’m really excited to apply more machine learning to this data, so that we can analyze it even faster,” says Tuxen-Bettman.

The maps and accompanying data are a boon to all concerned. Environmental scientists can take what’s been learned about mobile monitoring and hyperlocal pollution and apply it to cities and regions around the world with even worse air problems. On a more personal scale, individuals may someday be able to pull up an app on their smartphone and see whether it might be better for their lungs to walk down one street versus another a block over. For cities, the hyperlocal maps are a nice thing to have if you’re deciding on investing in a fleet of clean buses and just which areas might benefit from them the most. As for community organizations, the data is both confirmation and validation. “A lot of these areas that popped up on the map were areas that these community groups already knew were problems,” says Tuxen-Bettman. “They live there in the neighborhood. They see people getting sick. But these maps could finally prove what they have always suspected.”

ROBERT ITO is a writer based in Los Angeles. He is a frequent contributer to The New York Times, California Sunday, and Los Angeles magazine.

Photography by Cait Oppermann

Illustration by Brown Bird Design

Back to top