Utilizing Big Data for Built Environment Assessment: Alcohol, Food and Tobacco Retail Proximity

BACKGROUND: The built environment is the spaces and places created or modified by people that can encourage or discourage a variety of health behaviors. It is structured by urban design, economics and land use planning rules. The design and layout of communities can all be drivers of health by shaping our choices and options. In order to assess the built environment, public health needs to acquire and utilize large administrative datasets that are not traditionally meant for public health use. The Oregon Environmental Public Health Tracking Program (Oregon Tracking) has been utilizing a variety of data sources to assess the built environment – most recently, the proximity to alcohol, food and tobacco retailers.

METHODS: Oregon Tracking obtained the Department of Motor Vehicles (DMV) database and geocoded all addresses associated with state-issued drivers’ licenses and ID cards to pinpoint the location of Oregon households. Food retailer data was obtained via the Employment Department database and retail food types were classified using definitions from the North American Industry Classification System (NAICS). Alcohol retailers were identified using licensee data from the Oregon Liquor Control Commission (OLCC). Oregon Tracking examined three levels of OLCC business licenses; bars and restaurants, liquor stores, and other retailers. Tobacco retail outlets were identified using data from OLCC.

RESULTS: Oregon Tracking was able to get a near a complete listing of all Oregon households, food retailers and alcohol retailers. Not all tobacco retail outlets were included in the OLCC licensee database, which resulted in an undercount. Since the household and retail data was at the address level, Oregon Tracking was able to accurately calculate the following proximity and density measures - average walk distance, number per 1,000 population, number per square mile, and percent of households within ½ mile.

CONCLUSIONS: The DMV, Employment Department and the OLCC databases are inexpensive, reliable, stable and high quality datasets to utilize in the assessment of built environment measures as they pertain to alcohol, food and tobacco retail proximity. It is undetermined whether or not people make healthier choices based on proximity to these retail sources, but we are able to ascertain who lives in areas with an abundance or a lack of these retail outlets. These data can also assist in developing innovative strategies by providing data on the density of stores and their proximity to schools, to further help influence health corner stores movement, or even influence new land use regulations.