Capturing park-use behavior patterns using volunteered street view imagery: Does the likelihood increase with the volume of contributions

Abstract. Analyzing how citizens experience urban parks can provide valuable information for evaluating the effectiveness of park planning and management decisions to encourage park use. Although crowd-sourced data from social media have been reported to be useful for reflecting park visitations, limited user information from social media posts makes it challenging to understand the detailed characteristics of visitor behavior. To address this issue, this study evaluates the reliability of volunteered street view imagery (VSVI), a dataset of geo-tagged street-level images containing motion information from volunteers, as a proxy for park-use behavior, using metropolitan parks in the 23-ward of Tokyo as a case study. Several indicators of park-use behavior, including park visitor count, spatio-temporal distribution, and path characteristics were calculated using VSVI data and reference datasets to investigate data reliability. The results suggest that VSVI may serve as a complementary open dataset for identifying park-use patterns in parks with sufficient contribution volume, while also revealing contribution self-selection toward photography-oriented scenic or tourist parks. This study provides new insights for future research on modelling citizen activities using volunteered geographic information data.