Big Brother Meets Crowdsourcing
For help in tracking transportation and physical activity trends, Big Brother can come in handy.
Aaron Hipp, associate professor of community health and sustainability at NC State University, will show how to tap into publicly available webcam footage and use crowdsourcing to compile data on active transportation – walking and bicycling. He’s speaking at the Wireless Health 2015 meeting at the National Institutes of Health in Bethesda, Maryland, this week along with engineering colleague Veena Misra, one of the conference keynote presenters.
To get an accurate picture of walking, bicycling and driving patterns at selected sites in Washington, D.C., Hipp captured 17 hours of webcam images daily at each site for 17 months. He collected 29,000 webcam images from two intersections. To make sense of that mountain of data, he used Amazon Mechanical Turk – an online crowdsource – to find workers to count and categorize types of traffic for 2 cents an image. For reliability’s sake, four different MTurk workers marked the location of pedestrians, bicyclists and motorists in each image.
When trained research assistants did their own count, they found the crowdsourced results highly reliable.
Hipp, a faculty member in the College of Natural Resources and the Center for Geospatial Analytics, is interested in promoting health through walking and bicycling. He says webcam images and crowdsourced data can provide a source of unbiased, inexpensive information about how people actually travel each day. Most cities could tap into local images from traffic, red-light or speed cameras, making these methods highly scalable.
“The cameras are already out in public capturing daily life,” Hipp says. “Repurposing this data for public health surveillance gives us some insight into how to align changes in transportation policies and programs with changes in the environment.”
For details on Hipp’s Wireless Health 2015 presentation, read Cameras and Crowds in Transportation Tracking in the conference proceedings. The Built Environment and Public Health Clearinghouse published a related article, Using Webcams and Crowds to Study Active Transportation.