Leveraging 31 Million Google Street View Images to Characterize Built Environments and Examine County Health Outcomes

Abstract

Objectives Built environments can affect health, but data in many geographic areas are limited. We used a big data source to create national indicators of neighborhood quality and assess their associations with health. Methods We leveraged computer vision and Google Street View images accessed from December 15, 2017, through July 17, 2018, to detect features of the built environment (presence of a crosswalk, non–single-family home, single-lane roads, and visible utility wires) for 2916 US counties. We used multivariate linear regression models to determine associations between features of the built environment and county-level health outcomes (prevalence of adult obesity, prevalence of diabetes, physical inactivity, frequent physical and mental distress, poor or fair self-rated health, and premature death [in years of potential life lost]). Results Compared with counties with the least number of crosswalks, counties with the most crosswalks were associated with decreases of 1.3%, 2.7%, and 1.3% of adult obesity, physical inactivity, and fair or poor self-rated health, respectively, and 477 fewer years of potential life lost before age 75 (per 100 000 population). The presence of non–single-family homes was associated with lower levels of all health outcomes except for premature death. The presence of single-lane roads was associated with an increase in physical inactivity, frequent physical distress, and fair or poor self-rated health. Visible utility wires were associated with increases in adult obesity, diabetes, physical and mental distress, and fair or poor self-rated health. Conclusions The use of computer vision and big data image sources makes possible national studies of the built environment’s effects on health, producing data and results that may inform national and local decision-making.

Google Street View Images as Predictors of Patient Health Outcomes, 2017–2019

“…Uninsured patients were less likely to live in areas with green streets and non-single homes. In one of our previous studies, we found that greater county-level economic disadvantage was associated with a lower prevalence of non-single-family homes and visible wires at the county level after adjusting for violent crime rate, age, race/ethnicity, percentage of population not proficient in English, and ratio of population to primary care providers [ 49 ].…”

“…This is among the few studies examining GSV-derived predictors of individual-level outcomes, controlling for individual-level predisposing characteristics. Previous studies with GSV images have utilized ecological frameworks [ 48 ]: for instance, county-level built environment predictors of county health outcomes [ 49 ]. In partnership with one of the largest healthcare providers in Utah, in this study, we included close to one-third of the population in Utah.…”

Measuring Perceived Psychological Stress in Urban Built Environments Using Google Street View and Deep Learning

“…The emergence of crowdsourced mapping services and geotagged imagery containing a wealth of visual information [e.g., Baidu Street View, Tencent Street View, and Google Street View (GSV)] provides a usable source of big data. Mapping services provide academics and researchers with an application programming interface (API) for extracting high-spatial resolution images of streets and communities to reflect a built environment as its residents see it ( 14 ). GSV is now available in many countries and cities around the world, becoming a scholarly tool for studying the built environment of cities because of its wide coverage and high accuracy.…”

Machine Learning Approaches for Measuring Neighborhood Environments in Epidemiologic Studies

“…This approach typically requires mass downloading of Street View images in order to efficiently process the images on high-performance computational clusters. There have been notable successes in training machines to quantify trees and green space and to identify traffic control signs, crosswalks, single-lane roads, and utility wires in Street View panoramas [ 20 – 26 ]. Recent work has also used health data from surveys to train a generative adversarial networks (GAN) machine learning model to identify health-related features in Street View images [ 27 ].…”