Evaluating the influence of park attributes on user satisfaction from Flickr data for Southeast Michigan

The use of social media data is widely used by academics for research, including landscape architecture. The increased use of social media for expressing opinions about public open spaces presents opportunities to analyze this data from social media to evaluate park user satisfaction. Here we propose the following research question: Is there a correlation between the park attributes and the sentiments of visitors towards the landscape, and if so, can this data be used to inform future designs? To answer this question we use data from Flickr, a popular social media platform, as it offers the most viable data for landscape research since photos with attached descriptive titles or tags contain the geographic information which can be presented in geographic applications like ArcGIS, and it has no limitation of sharing duration. To conduct the research data from Flickr was accessed via an API. The subsequent data was cleaned, processed, and analyzed via several linear regression models. The results indicate some attributes of the parks have a strong relationship with user satisfaction: Golf course, Hunting Trapping Area, Natural Area, and Monument Historical Features. The landscape designers can focus more on those important attributes as a design guideline for the future, since the result of all the regression models indicates that those attributes influence significantly on user satisfaction.Master of Landscape Architecture (MLA)School for Environment and SustainabilityUniversity of Michiganhttp://deepblue.lib.umich.edu/bitstream/2027.42/174128/1/Yang, Jieping_Practicum.pd