Background
Physical activity plays a role in numerous health outcomes including obesity, diabetes, heart disease, and cancer. Over 30% of adults and 17% of children and adolescents in the US are obese, with lack of physical activity due to constraints in the built environment being an important influence. Lack of safe places to walk and bicycle and lack of access to parks and open space can impact the frequency, duration, and quality of physical activity of residents in urban settings. Physical activity may be purposive such as a jog in a park, or incidental such as a ten minute walk from home to a public transit stop. In both purposive and incidental cases the designs of urban built environments influence the decisions and experience of physical activity behaviors.

Objectives
Our team is investigating a line of research using publically available outdoor webcams, such as street intersection webcams, to capture active transportation in urban settings. A necessary initial step in this work is understanding the prevalence of active transportation across a variety of captured webcams.

Methods
Two webcams in the Archive of Many Outdoor Scenes (AMOS) captured the addition of a painted crosswalk (November 2007) for a commercial and residential street intersection in Washington, DC. For this analysis, we used photographs from AMOS captured every 30 minutes over a 14-month period between 7am and 7pm in both locations, before and after crosswalk additions (May-November, 2007 and 2008). The use of this webcam data allowed for a pre–post crosswalk travel-mode analysis across intersections located in different land use areas. Amazon Mechanical Turk (MTurk) was used to crowdsource image annotation, counting the number of pedestrians, cyclists, and vehicles per image. The odds of observing each transportation mode in Year 2 compared to Year 1 were examined. We are currently analyzing seasonal differences in peak active transportation in both locations.

Results
A total of 12076 pedestrians and 833 cyclists were observed in the commercial intersection, compared to 506 pedestrians and 166 cyclists in the residential location. The presence of a painted crosswalk predicted a significant increase in the number of pedestrians in both commercial (OR=1.62, 95% CI=1.36-1.93) and residential (OR=1.47, 95% CI=1.07-2.01) locations on weekdays. Pedestrian activity peaked at three times downtown (9am, 1pm, 6pm) and twice in the residential location (8am and 6pm) Afternoon biking activity peaked an hour earlier (6pm) in the downtown location than the residential location (7pm). Pedestrian activity was highest on Wednesdays for both groups. Both locations experienced the highest rates of bicycling and pedestrian activity during the summer season and on weekdays. Findings are consistent with previous observation and personnel-intensive studies of peak commuting activity.

Conclusions
Findings suggest webcams and crowdsourcing have great potential for capturing active transportation patterns. The use of public webcams and MTurks offer an inexpensive (US$0.02/photo) means to evaluate patterns of commuting behavior and potentially the effectiveness of built environment policies and interventions.

Implications
Using an eight-year archive of captured webcam images and crowdsources, we have demonstrated that improvements in urban built environments are associated with subsequent and significant increases in physical activity behaviors. Webcams are able to capture a variety of built environment attributes and our previous studies have shown that webcams are a reliable and valid source of built environment information. As such, the emerging technology of publicly available webcams facilitates both consistent uptake and potentially timely dissemination of physical activity and built environment behaviors across a variety of outdoor environments. The AMOS webcams have the potential to serve as an important and cost-effective part of urban environment and public health surveillance to evaluate patterns and trends of population-level physical activity behavior in diverse built environments.

References

1. Brownson, R. C., Hoehner, C. M., Day, K., Forsyth, A., & Sallis, J.F. . (2009). Measuring the Built Environment for Physical Activity: State of the Science. American Journal of Preventive Medicine, 36(4 Supplement), S99-123.e112. doi: 10.1016/j.amepre.2009.01.005
2. Jackson, R. J. (2003). The Impact of the Built Environment on Health: An Emerging Field. Am J Public Health, 93(9), 1382-1384. doi: 10.2105/AJPH.93.9.1382
3. Jackson, R. J., Dannenberg, Andrew L., & Frumkin, Howard. (2013). Health and the Built Environment: 10 Years After. American Journal of Public Health, 103(9), 1542-1544. doi: 10.2105/ajph.2013.301482
4. CDC. (2009). Division of Nutrition, Physical Activity and Obesity. Available from: http://www.cdc.gov/nccdphp/dnpa/index.htm.
5. CDC. (2011). Guide To Community Preventive Services. Atlanta, GA: Epidemiology Program Office, CDC.

Support / Funding Source
This work is supported by a National Cancer Institute (NCI) grant #1R21CA186481-01s. The opinions or assertions contained herein are the private ones of the authors and are not considered as official or reflecting the views of the NCI.