Abstract
The capabilities offered by smart sensing (the Internet of Things) and data science, create new opportunities to carry out large-scale studies involving social science and human factors. We report here our findings on a pilot study aimed at better understanding how citizens interact with urban green areas, identify relevant features, spot interaction patterns and, ultimately, recommend interventions to increase well-being. Our study was carried out in Sheffield (UK), where we tracked 1,870 subjects for two different periods (7 and 30 days), covering 760 digitally geo-fenced green areas. Through a smartphone App, we collected both subjective data (personal feelings, type of social interactions, type of activity, and perception of space) and objective data (sensor data, location, time, and photos). We employed data science methods to filter, correlate, cluster, and visualize the data, doing text analysis to extract semantic information from the subjects' responses. Looking at the intensity of interaction between citizens and green spaces, we found a stronger correlation with the quality of the green areas (diversity of natural features, trees, and birds), rather than their size (half of the top visited areas included small squares and gardens). Looking at the type of social interaction taking place within the green areas (lone visits or with friends and family), we found that different social interaction patterns correlate to different types of green area. Interestingly, most of the interactions correlate strongly to the proximity to the city centre, the presence of facilities (sport, parking), and architectural features (listed building, artistic/monumental icons), which indicate that targeted small-scale intervention into central built areas may have more immediate impact towards the citizen's well-being than the more peripheral parks.