Utility of passive photography to objectively audit built environment features of active transport journeys: an observational study

Using Computer Vision to Annotate Video-Recoded Direct Observation of Physical Behavior

Direct observation is a ground-truth measure for physical behavior, but the high cost limits widespread use. The purpose of this study was to develop and test machine learning methods to recognize aspects of physical behavior and location from videos of human movement: Adults (N = 26, aged 18-59 y) were recorded in their natural environment for two, 2- to 3-h sessions. Trained research assistants annotated videos using commercially available software including the following taxonomies: (1) sedentary versus non-sedentary (two classes); (2) activity type (four classes: sedentary, walking, running, and mixed movement); and (3) activity intensity (four classes: sedentary, light, moderate, and vigorous). Four machine learning approaches were trained and evaluated for each taxonomy. Models were trained on 80% of the videos, validated on 10%, and final accuracy is reported on the remaining 10% of the videos not used in training. Overall accuracy was as follows: 87.4% for Taxonomy 1, 63.1% for Taxonomy 2, and 68.6% for Taxonomy 3. This study shows it is possible to use computer vision to annotate aspects of physical behavior, speeding up the time and reducing labor required for direct observation. Future research should test these machine learning models on larger, independent datasets and take advantage of analysis of video fragments, rather than individual still images.

Automated wearable cameras for improving recall of diet and time use in Uganda: a cross-sectional feasibility study

BACKGROUND

Traditional recall approaches of data collection for assessing dietary intake and time use are prone to recall bias. Studies in high- and middle-income countries show that automated wearable cameras are a promising method for collecting objective health behavior data and may improve study participants' recall of foods consumed and daily activities performed. This study aimed to evaluate the feasibility of using automated wearable cameras in rural Eastern Ugandan to collect dietary and time use data.

METHODS

Mothers of young children (n = 211) wore an automated wearable camera on 2 non-consecutive days while continuing their usual activities. The day after wearing the camera, participants' dietary diversity and time use was assessed using an image-assisted recall. Their experiences of the method were assessed via a questionnaire.

RESULTS

Most study participants reported their experiences with the automated wearable camera and image-assisted recall to be good (36%) or very good (56%) and would participate in a similar study in the future (97%). None of the eight study withdrawals could be definitively attributed to the camera. Fifteen percent of data was lost due to device malfunction, and twelve percent of the images were "uncodable" due to insufficient lighting. Processing and analyzing the images were labor-intensive, time-consuming, and prone to human error. Half (53%) of participants had difficulty interpreting the images captured by the camera.

CONCLUSIONS

Using an automated wearable camera in rural Eastern Uganda was feasible, although improvements are needed to overcome the challenges common to rural, low-income country contexts and reduce the burdens posed on both participants and researchers. To improve the quality of data obtained, future automated wearable camera-based image assisted recall studies should use a structured data format to reduce image coding time; electronically code the data in the field, as an output of the image review process, to eliminate ex post facto data entry; and, ideally, use computer-assisted personal interviews software to ensure completion and reduce errors. In-depth formative work in partnership with key local stakeholders (e.g., researchers from low-income countries, representatives from government and/or other institutional review boards, and community representatives and local leaders) is also needed to identify practical approaches to ensuring that the ethical rights of automated wearable camera study participants in low-income countries are adequately protected.

GeoAI for Large-Scale Image Analysis and Machine Vision: Recent Progress of Artificial Intelligence in Geography

GeoAI, or geospatial artificial intelligence, has become a trending topic and the frontier for spatial analytics in Geography. Although much progress has been made in exploring the integration of AI and Geography, there is yet no clear definition of GeoAI, its scope of research, or a broad discussion of how it enables new ways of problem solving across social and environmental sciences. This paper provides a comprehensive overview of GeoAI research used in large-scale image analysis, and its methodological foundation, most recent progress in geospatial applications, and comparative advantages over traditional methods. We organize this review of GeoAI research according to different kinds of image or structured data, including satellite and drone images, street views, and geo-scientific data, as well as their applications in a variety of image analysis and machine vision tasks. While different applications tend to use diverse types of data and models, we summarized six major strengths of GeoAI research, including (1) enablement of large-scale analytics; (2) automation; (3) high accuracy; (4) sensitivity in detecting subtle changes; (5) tolerance of noise in data; and (6) rapid technological advancement. As GeoAI remains a rapidly evolving field, we also describe current knowledge gaps and discuss future research directions.

Validation of an Automated Wearable Camera-Based Image-Assisted Recall Method and the 24-h Recall Method for Assessing Women's Time Allocation in a Nutritionally Vulnerable Population: The Case of Rural Uganda

Accurate data are essential for investigating relationships between maternal time-use patterns and nutritional outcomes. The 24 h recall (24HR) has traditionally been used to collect time-use data, however, automated wearable cameras (AWCs) with an image-assisted recall (IAR) may reduce recall bias. This study aimed to evaluate their concurrent criterion validity for assessing women’s time use in rural Eastern Ugandan. Women’s (n = 211) time allocations estimated via the AWC-IAR and 24HR methods were compared with direct observation (criterion method) using the Bland–Altman limits of agreement (LOA) method of analysis and Cronbach’s coefficient alpha (time allocation) or Cohen’s κ (concurrent activities). Systematic bias varied from 1 min (domestic chores) to 226 min (caregiving) for 24HR and 1 min (own production) to 109 min (socializing) for AWC-IAR. The LOAs were within 2 h for employment, own production, and self-care for 24HR and AWC-IAR but exceeded 11 h (24HR) and 9 h (AWC-IAR) for caregiving and socializing. The LOAs were within four concurrent activities for 24HR (−1.1 to 3.7) and AWC-IAR (−3.2 to 3.2). Cronbach’s alpha for time allocation ranged from 0.1728 (socializing) to 0.8056 (own production) for 24HR and 0.2270 (socializing) to 0.7938 (own production) for AWC-IAR. For assessing women’s time allocations at the population level, the 24HR and AWC-IAR methods are accurate and reliable for employment, own production, and domestic chores but poor for caregiving and socializing. The results of this study suggest the need to revisit previously published research investigating the associations between women’s time allocations and nutrition outcomes.

Objective measurement of children's physical activity geographies: A systematic search and scoping review

This study aimed to systematically identify, map out, and describe geographical information systems (GIS)-based approaches that have been employed to measure children's neighborhood geographies for physical activity behaviors. Forty studies were included, most were conducted in the USA. Heterogeneity in GIS methods and measures was found. The majority of studies estimated children's environments using Euclidean or network buffers ranging from 100 m to 5 km. No singular approach to measuring children's physical activity geographies was identified as optimal. Geographic diversity in studies as well as increased use of measures of actual neighborhood exposure are needed. Improved consistency and transparency in reporting research methods is urgently required.

•

Varied GIS measures of children's physical activity geographies were identified.

•

Evidence was heterogeneous and predominantly from the USA.

•

Most research used Euclidean or network buffers ranging from 100 m to 5 km.

•

Larger buffer sizes (i.e., ≥800 m) performed better than smaller buffer sizes.

•

No optimal approach to measuring children's activity geographies was determined.

Derivation of Time-Activity Data Using Wearable Cameras and Measures of Personal Inhalation Exposure among Workers at an Informal Electronic-Waste Recovery Site in Ghana

OBJECTIVES

Approximately 2 billion workers globally are employed in informal settings, which are characterized by substantial risk from hazardous exposures and varying job tasks and schedules. Existing methods for identifying occupational hazards must be adapted for unregulated and challenging work environments. We designed and applied a method for objectively deriving time-activity patterns from wearable camera data and matched images with continuous measurements of personal inhalation exposure to size-specific particulate matter (PM) among workers at an informal electronic-waste (e-waste) recovery site.

METHODS

One hundred and forty-two workers at the Agbogbloshie e-waste site in Accra, Ghana, wore sampling backpacks equipped with wearable cameras and real-time particle monitors during a total of 171 shifts. Self-reported recall of time-activity (30-min resolution) was collected during the end of shift interviews. Images (N = 35,588) and simultaneously measured PM2.5 were collected each minute and processed to identify activities established through worker interviews, observation, and existing literature. Descriptive statistics were generated for activity types, frequencies, and associated PM2.5 exposures. A kappa statistic measured agreement between self-reported and image-based time-activity data.

RESULTS

Based on image-based time-activity patterns, workers primarily dismantled, sorted/loaded, burned, and transported e-waste materials for metal recovery with high variability in activity duration. Image-based and self-reported time-activity data had poor agreement (kappa = 0.17). Most measured exposures (90%) exceeded the World Health Organization (WHO) 24-h ambient PM2.5 target of 25 µg m-3. The average on-site PM2.5 was 81 µg m-3 (SD: 94). PM2.5 levels were highest during burning, sorting/loading and dismantling (203, 89, 83 µg m-3, respectively). PM2.5 exposure during long periods of non-work-related activities also exceeded the WHO standard in 88% of measured data.

CONCLUSIONS

In complex, informal work environments, wearable cameras can improve occupational exposure assessments and, in conjunction with monitoring equipment, identify activities associated with high exposures to workplace hazards by providing high-resolution time-activity data.

Out-of-Home Mobility and Social Participation of Older People: a Photo-Based Ambulatory Assessment Study

Research has shown that social participation contributes to older people’s healthy ageing. Out-of-home mobility may promote higher levels of social participation among older people. However, mobility and social participation are sophisticated constructs. Social participation can be distinguished among different levels, such as being alone, being alone surrounded by others, interacting with others or engaging in activities together. Social participation and time spent in out-of-home-mobility can be difficult to remember and therefore difficult to assess. Picture-based ambulatory assessment provides valid and reliable information about people’s mobility as well as their level of participation with high ecological validity. The aim of the study was to investigate older people’s level of social participation and its association with high or low out-of-home mobility. In sum, 23072 pictures (mean per person = 2307; SD = 686.7) involving ten older people (mean age = 75.4 years; SD = 7.5 years) living in southwestern Germany were analysed. They were asked to wear a wearable camera for two consecutive days. Images were automatically captured every 15 seconds. Image analysis shows that study participants spent most of their time alone (at approximately 35% of analysed time). Out-of-home mobility was associated with higher levels of social participation, such as helping others (chi2 = 200,664, df = 5, p < .001). Picture-based ambulatory assessment can assist in the gathering of necessary sophisticated information that is difficult to assess via questionnaires or other similar methods.

Daily metabolic expenditures: estimates from US, UK and polish time-use data

BACKGROUND

Behaviour has diverse economic, social and health consequences. Linking time spent in different daily activities to energy expenditure (EE) is one way of investigating the health and physiological consequences of behaviour and identifying targets to improve population health and well-being.

METHODS

We estimated behaviour-related EE for respondents to time use surveys (TUS) from three countries: UK 2001, Poland 2012 and US 2003-13. The Harmonised Multinational Time Use Survey (MTUS) activity categories were matched to MET estimates from the 2011 Compendium of Physical Activities. We attach METs values to each successive activity in the TUS, together with both the original UK, Polish and US activity classifications and the 68-category MTUS activity classification. We used TUS estimates of activity durations across 24-h to estimate the Physical Activity Level (PAL) for respondents from the three countries and the average time spent and MET values for different activity categories.

RESULTS

PAL values ranged from 1.59 in the US to 1.74 in Poland. The main sources of daily EE from PA were paid and unpaid work activities. Discretionary PA accounted for only a very small part (~ 3%) of adult daily energy expenditures. Using the harmonised MTUS 68-activity classification reduced the variability of the aggregate PAEE measure by ~ 20%, but the patterns of association between key demographics (age, sex, educational attainment) were unaffected. TUS data were further used to (1) identify sources of daily PA, and (2) assess adherence to physical activity guidelines (PAG) on a single-day basis. Estimated adherence levels were similar to those reported from other TUS as well as frequency based estimates.

CONCLUSIONS

Comparative studies of energy expenditure based on harmonised time use activity categories could provide insight into the relative importance of different activities for energy expenditure across different countries and demographic groups. However, new observational studies combining TUS data with accelerometer, direct observation and other measures of activity intensity are required for more accurate MET assignments to activity categories in TUS.

Feasibility of wearable cameras to assess screen time and time spent restrained in children aged 3 to 5 years: a study protocol

INTRODUCTION

Wearable cameras may help overcome the limitations of existing tools to measure young children's sedentary behaviour, but their use introduces a range of ethical challenges. The primary aim of this study is to determine the feasibility of using wearable cameras to measure the two aspects of sedentary behaviour currently included in evidence-based guidelines (ie, screen time and time spent restrained). If shown to be feasible, a secondary aim will be to validate subjective measures against the directly measured screen time and time spent restrained.

METHODS AND ANALYSIS

A convenience sample (n=20) will be recruited via flyers at the University of Strathclyde and advertisements on online forums for parents of young children (aged 3 to 5 years). Children will be provided with a wearable camera, attached to the front of their clothing with a clip, to be worn for 3 days (2 non-childcare days and 1 weekend day) in non-public settings. Once switched on, the camera will record continuous video footage. Parents will complete an online survey providing their feedback on their own and their child's experience of the wearable camera. They will also report their own and their child's demographical characteristics and their child's usual daily screen time and time spent restrained in the past week. Data will be downloaded using specialised software and second-by-second coding will be undertaken. Feasibility and acceptability will be assessed using percentages and by analysing qualitative data. If feasibility is shown, intraclass coefficients will be used to determine agreement between video data and parent-reported sedentary behaviours.

ETHICS AND DISSEMINATION

Ethical approval has been granted by the School of Psychological Sciences and Health Ethics Committee at the University of Strathclyde. Results will be used to inform future studies and will be disseminated in peer-reviewed journals and at major international conferences.

Can Wearable Cameras be Used to Validate School-Aged Children's Lifestyle Behaviours?

Wearable cameras combined with accelerometers have been used to estimate the accuracy of children's self-report of physical activity, health-related behaviours, and the contexts in which they occur. There were two aims to this study; the first was to validate questions regarding self-reported health and lifestyle behaviours in 9⁻11-year-old children using the child's health and activity tool (CHAT), an accelerometer and a wearable camera. Second, the study sought to evaluate ethical challenges associated with taking regular photographs using a wearable camera through interviews with children and their families. Fourteen children wore an autographer and hip-worn triaxial accelerometer for the waking hours of one school and one weekend day. For both of these days, children self-reported their behaviours chronologically and sequentially using the CHAT. Data were examined using limits of agreement and percentage agreement to verify if reference methods aligned with self-reported behaviours. Six parent⁻child dyads participated in interviews. Seven, five, and nine items demonstrated good, acceptable, and poor validity, respectively. This demonstrates that the accuracy of children's recall varies according to the behaviour or item being measured. This is the first study to trial the use of wearable cameras in assessing the concurrent validity of children's physical activity and behaviour recall, as almost all other studies have used parent proxy reports alongside accelerometers. Wearable cameras carry some ethical and technical challenges, which were examined in this study. Parents and children reported that the autographer was burdensome and in a few cases invaded privacy. This study demonstrates the importance of adhering to an ethical framework.

Advances and Controversies in Diet and Physical Activity Measurement in Youth

Technological advancements in the past decades have improved dietary intake and physical activity measurements. This report reviews current developments in dietary intake and physical activity assessment in youth. Dietary intake assessment has relied predominantly on self-report or image-based methods to measure key aspects of dietary intake (e.g., food types, portion size, eating occasion), which are prone to notable methodologic (e.g., recall bias) and logistic (e.g., participant and researcher burden) challenges. Although there have been improvements in automatic eating detection, artificial intelligence, and sensor-based technologies, participant input is often needed to verify food categories and portions. Current physical activity assessment methods, including self-report, direct observation, and wearable devices, provide researchers with reliable estimations for energy expenditure and bodily movement. Recent developments in algorithms that incorporate signals from multiple sensors and technology-augmented self-reporting methods have shown preliminary efficacy in measuring specific types of activity patterns and relevant contextual information. However, challenges in detecting resistance (e.g., in resistance training, weight lifting), prolonged physical activity monitoring, and algorithm (non)equivalence remain to be addressed. In summary, although dietary intake assessment methods have yet to achieve the same validity and reliability as physical activity measurement, recent developments in wearable technologies in both arenas have the potential to improve current assessment methods.

THEME INFORMATION

This article is part of a theme issue entitled Innovative Tools for Assessing Diet and Physical Activity for Health Promotion, which is sponsored by the North American branch of the International Life Sciences Institute.

Wearable camera-derived microenvironments in relation to personal exposure to PM2.5

Data regarding which microenvironments drive exposure to air pollution in low and middle income countries are scarce. Our objective was to identify sources of time-resolved personal PM_2.5 exposure in peri-urban India using wearable camera-derived microenvironmental information. We conducted a panel study with up to 6 repeated non-consecutive 24 h measurements on 45 participants (186 participant-days). Camera images were manually annotated to derive visual concepts indicative of microenvironments and activities. Men had slightly higher daily mean PM_2.5 exposure (43 μg/m³) compared to women (39 μg/m³). Cameras helped identify that men also had higher exposures when near a biomass cooking unit (mean (sd) μg/m³: 119 (383) for men vs 83 (196) for women) and presence in the kitchen (133 (311) for men vs 48 (94) for women). Visual concepts associated in regression analysis with higher 5-minute PM_2.5 for both sexes included: smoking (+93% (95% confidence interval: 63%, 129%) in men, +29% (95% CI: 2%, 63%) in women), biomass cooking unit (+57% (95% CI: 28%, 93%) in men, +69% (95% CI: 48%, 93%) in women), visible flame or smoke (+90% (95% CI: 48%, 144%) in men, +39% (95% CI: 6%, 83%) in women), and presence in the kitchen (+49% (95% CI: 27%, 75%) in men, +14% (95% CI: 7%, 20%) in women). Our results indicate wearable cameras can provide objective, high time-resolution microenvironmental data useful for identifying peak exposures and providing insights not evident using standard self-reported time-activity.

Engaging research participants to inform the ethical conduct of mobile imaging, pervasive sensing, and location tracking research

Researchers utilize mobile imaging, pervasive sensing, social media, and location tracking (MISST) technologies to observe and intervene with participants in their natural environment. The use of MISST methods and tools introduces unique ethical issues due to the type and quantity of data, and produces raising new challenges around informed consent, risk assessment, and data management. Since MISST methods are relatively new in behavioral research, there is little documented evidence to guide institutional review board (IRB) risk assessment and inform appropriate risk management strategies. This study was conducted to contribute the participant perspectives when considering ethical and responsible practices. Participants (n = 82) enrolled in an observational study where they wore several MISST devices for 1 week completed an exit survey. Survey items focused on the following: 1-device comfort, 2-informed consent, 3-privacy protections, and 4-bystander engagement. The informed consent process reflected participant actual experience. Device comfort and privacy were raised as concerns to both the participants and bystanders. While the majority of the participants reported a positive experience, it is important to note that the participants were volunteers who were not mandated to wear tracking devices and that persons who are mandated may not have a similar response. Findings support strategies proposed in the Kelly et al. (2013) ethical framework, which emphasizes procedures to improve informed consent, protect privacy, manage data, and respect bystander rights when using a wearable camera.

Kids in space: Measuring children's residential neighborhoods and other destinations using activity space GPS and wearable camera data

INTRODUCTION

Defining the boundary of children's 'neighborhoods' has important implications for understanding the contextual influences on child health. Additionally, insight into activities that occur outside people's neighborhoods may indicate exposures that place-based studies cannot detect. This study aimed to 1) extend current neighborhood research, using data from wearable cameras and GPS devices that were worn over several days in an urban setting; 2) define the boundary of children's neighborhoods by using leisure time activity space data; and 3) determine the destinations visited by children in their leisure time, outside their neighborhoods.

METHOD

One hundred and fourteen children (mean age 12y) from Wellington, New Zealand wore wearable cameras and GPS recorders. Residential Euclidean buffers at incremental distances were paired with GPS data (thereby identifying time spent in different places) to explore alternative definitions of neighborhood boundaries. Children's neighborhood boundary was at 500 m. A newly developed software application was used to identify 'destinations' visited outside the neighborhood by specifying space-time parameters. Image data from wearable cameras were used to determine the type of destination.

RESULTS

Children spent over half of their leisure time within 500 m of their homes. Children left their neighborhood predominantly to visit school (for leisure purposes), other residential locations (e.g. to visit friends) and food retail outlets (e.g. convenience stores, fast food outlets). Children spent more time at food retail outlets than at structured sport and in outdoor recreation locations combined.

CONCLUSION

Person-centered neighborhood definitions may serve to better represent children's everyday experiences and neighborhood exposures than previous methods based on place-based measures. As schools and other residential locations (friends and family) are important destinations outside the neighborhood, such destinations should be taken into account. The combination of image data and activity space GPS data provides a more robust approach to understanding children's neighborhoods and activity spaces.

Kids'Cam: An Objective Methodology to Study the World in Which Children Live

INTRODUCTION

This paper reports on a new methodology to objectively study the world in which children live. The primary research study (Kids'Cam Food Marketing) illustrates the method; numerous ancillary studies include exploration of children's exposure to alcohol, smoking, "blue" space and gambling, and their use of "green" space, transport, and sun protection.

METHODS

One hundred sixty-eight randomly selected children (aged 11-13 years) recruited from 16 randomly selected schools in Wellington, New Zealand used wearable cameras and GPS units for 4 days, recording imagery every 7 seconds and longitude/latitude locations every 5 seconds. Data were collected from July 2014 to June 2015. Analysis commenced in 2015 and is ongoing. Bespoke software was used to manually code images for variables of interest including setting, marketing media, and product category to produce variables for statistical analysis. GPS data were extracted and cleaned in ArcGIS, version 10.3 for exposure spatial analysis.

RESULTS

Approximately 1.4 million images and 2.2 million GPS coordinates were generated (most were usable) from many settings including the difficult to measure aspects of exposures in the home, at school, and during leisure time. The method is ethical, legal, and acceptable to children and the wider community.

CONCLUSIONS

This methodology enabled objective analysis of the world in which children live. The main arm examined the frequency and nature of children's exposure to food and beverage marketing and provided data on difficult to measure settings. The methodology will likely generate robust evidence facilitating more effective policymaking to address numerous public health concerns.

Technologies That Assess the Location of Physical Activity and Sedentary Behavior: A Systematic Review

Background

The location in which physical activity and sedentary behavior are performed can provide valuable behavioral information, both in isolation and synergistically with other areas of physical activity and sedentary behavior research. Global positioning systems (GPS) have been used in physical activity research to identify outdoor location; however, while GPS can receive signals in certain indoor environments, it is not able to provide room- or subroom-level location. On average, adults spend a high proportion of their time indoors. A measure of indoor location would, therefore, provide valuable behavioral information.

Objective

This systematic review sought to identify and critique technology which has been or could be used to assess the location of physical activity and sedentary behavior.

Methods

To identify published research papers, four electronic databases were searched using key terms built around behavior, technology, and location. To be eligible for inclusion, papers were required to be published in English and describe a wearable or portable technology or device capable of measuring location. Searches were performed up to February 4, 2015. This was supplemented by backward and forward reference searching. In an attempt to include novel devices which may not yet have made their way into the published research, searches were also performed using three Internet search engines. Specialized software was used to download search results and thus mitigate the potential pitfalls of changing search algorithms.

Results

A total of 188 research papers met the inclusion criteria. Global positioning systems were the most widely used location technology in the published research, followed by wearable cameras, and radio-frequency identification. Internet search engines identified 81 global positioning systems, 35 real-time locating systems, and 21 wearable cameras. Real-time locating systems determine the indoor location of a wearable tag via the known location of reference nodes. Although the type of reference node and location determination method varies between manufacturers, Wi-Fi appears to be the most popular method.

Conclusions

The addition of location information to existing measures of physical activity and sedentary behavior will provide important behavioral information.

Assessing environmental features related to mental health: a reliability study of visual streetscape images

BACKGROUND

An association between depressive symptoms and features of built environment has been reported in the literature. A remaining research challenge is the development of methods to efficiently capture pertinent environmental features in relevant study settings. Visual streetscape images have been used to replace traditional physical audits and directly observe the built environment of communities. The aim of this work is to examine the inter-method reliability of the two audit methods for assessing community environments with a specific focus on physical features related to mental health.

METHODS

Forty-eight postcodes in urban and rural areas of Cambridgeshire, England were randomly selected from an alphabetical list of streets hosted on a UK property website. The assessment was conducted in July and August 2012 by both physical and visual image audits based on the items in Residential Environment Assessment Tool (REAT), an observational instrument targeting the micro-scale environmental features related to mental health in UK postcodes. The assessor used the images of Google Street View and virtually "walked through" the streets to conduct the property and street level assessments. Gwet's AC1 coefficients and Bland-Altman plots were used to compare the concordance of two audits.

RESULTS

The results of conducting the REAT by visual image audits generally correspond to direct observations. More variations were found in property level items regarding physical incivilities, with broad limits of agreement which importantly lead to most of the variation in the overall REAT score. Postcodes in urban areas had lower consistency between the two methods than rural areas.

CONCLUSIONS

Google Street View has the potential to assess environmental features related to mental health with fair reliability and provide a less resource intense method of assessing community environments than physical audits.

Using Google Street View for systematic observation of the built environment: analysis of spatio-temporal instability of imagery dates

BACKGROUND

Recently, Google Street View (GSV) has been examined as a tool for remotely conducting systematic observation of the built environment. Studies have found it offers benefits over in-person audits, including efficiency, safety, cost, and the potential to expand built environment research to larger areas and more places globally. However, one limitation has been the lack of documentation on the date of imagery collection. In 2011, Google began placing a date stamp on images which now enables investigation of this concern. This study questions the spatio-temporal stability in the GSV date stamp. Specifically, is the imagery collected contemporaneously? If not, how frequently and where is imagery from different time periods woven together to represent environmental conditions in a particular place. Furthermore, how much continuity exists in imagery for a particular time period? Answering these questions will provide guidance on the use of GSV as a tool for built environment audits.

METHODS

GSV was used to virtually "drive" five sites that are a part of the authors' ongoing studies. Each street in the sites was "driven" one mouse-click at a time while observing the date stamp on each image. Every time the date stamp changed, this "disruption" was marked on the map. Every street segment in the site was coded by the date the imagery for that segment was collected. Spatial query and descriptive statistics were applied to understand the spatio-temporal patterns of imagery dates.

RESULTS

Spatio-temporal instability is present in the dates of GSV imagery. Of the 353 disruptions, 82.4% occur close to (<25 m) intersections. The remainder occurs inconsistently in other locations. The extent of continuity for a set of images collected with the same date stamp ranged from 3.13 m to 3373.06 m, though the majority of continuous segments were less than 400 m.

CONCLUSION

GSV offers some benefits over traditional built environment audits. However, this investigation empirically identifies a previously undocumented limitation in its application for research. Imagery dates can change often and without warning. Caution should be used at intersections where these disruptions are most likely to occur, though caution should be used everywhere when using GSV as a data collection tool.