Quantifying Biases in Social Media Analysis of Recreation in Urban Parks

Abstract

Recent years have seen an increase in the use of social media data for various decision-making purposes in the context of Urban Computing and smart cities, including management of public parks. Parks are critical natural assets that promote the health and well-being of urban residents, making it important that managers consider information about the quantity and character of park use in location and design decisions. However, as policy and management decisions rely on more autonomous methods, a critical concern that arises is the extent to which such analysis is fair and inclusive. In this article, we examine the biases that exist in data that are commonly used for the purpose of quantifying recreational use of urban parks. More precisely, we demonstrate the biases that exist in different sources of social media by comparing posts that are shared on Instagram and Flickr from ten urban parks in Seattle, WA. We compare the extent to which these platforms differ in terms of the information they capture about the number of people that visit the selected locations. We then demonstrate how further biases may be imposed when leveraging artificial intelligence to detect the count and demographics of park visitors, by comparing against an intercept survey of visitors.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 4 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.