Does geo-located social media reflect the visit frequency of urban parks? A city-wide analysis using the count and content of photographs

Abstract

Given the importance of parks and green spaces for outdoor recreation in cities, numerous studies have attempted to describe patterns of usage and understand their determining factors. Recently, social media has emerged as a potentially valuable tool to examine people’s use of parks. This study examines park use in Singapore based on the count and visual content of photographs geo-located within parks. Measures of park use—the number of photo-user-days (PUD)—derived from 325,173 and 94,890 photographs on the respective platforms Instagram and Flickr were compared with results from household surveys (n = 2000). We analysed the spatial attributes of parks and their relationships with PUD at an aggregated-level, and for content categories on the Flickr platform produced by automated classification: birds, wildlife, plants, flowers, recreation, water/skyscapes. In contrast to studies of large national parks, we found that PUD at city parks reflected residents’ preferences better than their frequency of visits to parks, and that park size had a limited effect on PUD. Some relationships were specific to a particular platform; Instagram users were more likely to upload photographs at parks that were closer to the coast and with more canopy cover, while Flickr users tended to do so at parks with an event space and that had lower-density housing nearby. We conclude that social media can provide reasonable assessments of park popularity, but future studies need to consider scale-effects, the integration of data sources for better accuracy, as well as a diversity of goals beyond park use.