Why Understanding the Big Data We Use for Research Matters – a Global Analysis of Flickr

Abstract Understanding the basic characteristics of a dataset is essential for making informed decisions on its suitability for research and beyond. Even simple descriptives, such as how much data there are overall and who produce it, are not easily acquired for big and user-generated data. We describe the spatio-temporal distribution of 229 million georeferenced posts by 1.5 million users in 2010–2022 on the photo-sharing platform Flickr. The platform’s popularity has declined, and increasingly the content is produced by a small group of super-users. Flickr is primarily used in the Global North. Additionally, we found that the platform’s public API did not respond consistently to spatially bounded queries. We conclude by discussing the possible impacts of these biases and data acquisition issues.