Crowdsourced air quality sensors provide opportunity to fill gaps left by the Environmental Protection Agency (EPA) in data resolution and availability by Sean Dasey
Fire and smoke erupted from an accident at the Chevron Refinery in Richmond, CA, in August 2012. Source: https://blogs.sfweekly.com/thesnitch/2012/08/fiery_explosion_rocks_east_bay.php
In 2014, the American Lung Association ranked Fresno, CA as the most polluted of 277 metropolitan areas in America for 24-hour particle pollution.  Salinas, CA was ranked the cleanest. Fresno and Salinas are only separated by 100 miles. This underscores the fact that air quality can vary greatly across nearby regions.
Although the current EPA air quality monitoring network captures differences across regions, it is not equipped to capture differences across localities and neighborhoods. For example, the 9-county San Francisco Bay Area, home to 7.5 million residents, has only 39 stations. In fact, this network failed to provide sufficiently detailed neighborhood scale air quality data during the August 6, 2012 Chevron refinery fire in Richmond, CA (pictured above), after which 11,000 nearby residents made emergency room visits for respiratory issues.  The closest EPA particulates monitor was located 2 miles away from the refinery, and the results took two weeks to analyze.
Historically, the EPA has focused on capturing air quality data on a regional scale by placing several monitoring stations per county, away from major roads or industrial emission sources in order to measure region-wide averages. The current EPA network of stations, however, is not designed to reveal how pollution sources affect the air quality of nearby neighborhoods with any real specificity.
Map of Bay Area District Air Monitor Sites; the dots on the map indicate where at least one full year of high quality wind speed, wind direction, and temperature data are archived that are suitable for modeling purposes. Source: https://hank.baaqmd.gov/tec/maps/dam_sites.htm#
One possible solution to this problem would be to crowdsource cheaper sensors to fill in the gaps between sporadically placed EPA stations.
In 2012, an open-source carbon monoxide and nitrogen dioxide sensor called the Air Quality Egg was successfully crowd funded on Kickstarter, and was named “Best of Kickstarter: 2012”. Although the Egg is not calibrated and therefore doesn’t give accurate readings on an individual basis the way EPA sensors do, it is several orders of magnitude cheaper than a calibrated sensor. This gives citizen scientists the opportunity to deploy them at a much larger scale than the existing EPA station network.
The Air Quality Egg began as a Kickstarter-funded sensor that is far less expensive and easier to use than traditional air quality monitors. Source: https://airqualityegg.com/
Louisville, KY, a historically industrial city that was once described as “smoky and blackened” by a visiting Charles Dickens, is emerging today as a leader in utilizing Eggs to monitor air quality on a neighborhood scale. On April 24, 2014, the nonprofit Institute for Healthy Air, Water and Soil, led by philanthropist Christy Brown and endorsed by Louisville Mayor Greg Fischer, announced the deployment of 100 Eggs around Louisville.  They will be deployed strategically, in neighborhoods downwind of heavily industrial zones.
The Egg is one example of a crowdsourced sensor contributing to an emerging network of sensors providing environmental and health data at the neighborhood scale. In Louisville, the Institute for Healthy Air, Water and Soil has partnered with the Louisville Asthma Data Innovation Project to correlate Egg data with time and location data from Bluetooth enabled asthma inhalers. As opposed to the traditional method of accumulating respiratory health data from county hospitals to get regional-scale statistics, this new method will lead to more precise location based data to pinpoint specific neighborhoods where air quality and respiratory health are concerns.
Hotspots of Athsma incidence in Louisville, KY. Source: https://www.slideshare.net/HealthDataConsortium/louisville-asthmapolis
Using RoundhouseOne’s proprietary data management system, 4Daptive, we can analyze correlations between data from the Egg we’ve deployed in Louisville and our larger sensor network, measuring thermal comfort, foot traffic, and outdoor weather conditions. 4Daptive provides the ability to manage data through an organized database, analyze correlations between multiple data sources, and produce user-created charts accompanied by customizable statistical outputs.
Limitations of available data have traditionally forced air quality analyses to be done on a regional scale. The emergence of localized, neighborhood-scale data will provide opportunities for new insight on how air quality affects our daily lives.
Sean Dasey is a Data Analyst at RoundhouseOne, MKThink’s in-house data analytics team. His work focuses on studying the effect of building design on thermal comfort, indoor air quality, and energy use.