Twelfth International Conference on The Image

Abstract

Since its creation in June 2020, the George Floyd and Anti-Racist Street Art database has collected photographs of over 2000 pieces of graffiti and street art created during and after the protests following the killing of George Floyd. When I had the opportunity to work with this archive, I noticed that the metadata was inconsistent because of the large number of people inputting items. I started a project to improve the metadata by analyzing the tags attached to images and attempting to regularize the vocabulary used so that images could be effectively organized and retrieved. A literature review into the organization of street art archives revealed concentration on collection methods, with much less attention to organization. It was thus necessary to construct a tailored tagging scheme based on generalized metadata best practices. The database was unsuitable for a controlled vocabulary, but an uncontrolled vocabulary had resulted in inability to accurately search the archive. My solution was to develop a semi-controlled vocabulary, where guidelines regularize usage of the more common tags and steer users into best practices. These tags enable encoding of not just the physical characteristics of images but also the content, thus textualizing the images for future retrieval. By creating didactics educating all stakeholders in the semi-controlled tagging vocabulary, I have enabled a process where the metadata should be regularized as it is entered, creating a useful dataset for researchers without requiring archivists to extensively re-edit the metadata. I believe this method would be adaptable to other image-based archives.