This novel interdisciplinary collaboration between public health and computer science provides automatic analysis of existing public data feeds to quantify the impact of built environment intervention on increasing bike travel mode share.
- The Archive of Many Outdoor Scenes (AMOS) has archived over 225 million images of outdoor environments from more than 12,000 public webcams since 2006. Using the publicly available webcams and a custom web crawler (similar to the web search engine or Google), webcam images are captured at the rate of one image per camera per hour and given a time stamp. Many of the locations have had built environment improvements such as complete streets, bike share startups or walking school bus programs. AMOS is able to document and allow quantification of population behavior changes following the built environment modification.
- The intersection of Pennsylvania Ave NW and 9th ST NW in Washington, DC where bike lanes were installed was chosen as a location to monitor transportation mode share comparing the first workweek or June 2009 and the first week of June 2010 (pre-bicycle lane and post-bicycle lane).
- Amazon Mechanical Turk (MTurk) website was used to crowd-source the image annotation. MTurks are simple tasks not yet computer automated. MTurk workers were paid US $0.01, in March 2012 to count each pedestrian, cyclist, and vehicle in a photograph. Each image was counted 5 unique times (n=1200), completed in less than 8 hours. The counts per transportation mode were downloaded to SPSSv.19 for analysis. Results showed a statistically significant difference in transportation mode share between the two years: no significant increase in pedestrians but a four-fold increase in the number of cyclists per scene.
- The investigators conclude that the use of AMOS and MTurks offer an inexpensive ($12.00 for this study) opportunity to quantify behavior change impact following built environment changes. Future plans include monitoring other locations in the Washington DC Capitol Bikeshare program and developing computer algorithms to automate the counting process.
Hipp, J. Aaron; Adlakha, Deepti; Chang, Bill; Eyler, Amy A.; and Pless, Robert B. (2013). "Emerging Technologies: Webcams and Crowd- Sourcing to Identify Active Transportation" (2013). Brown School Faculty Publications (Paper 3).