Image Activism After the Arab Uprisings| Challenges in Codifying Events Within Large and Diverse Data Sets of Human Rights Documentation: Memory, Intent, and Bias

Abstract

This article discusses challenges in codifying events within large and diverse data sets of human rights documentation, focusing on issues related to memory, intent, and bias. Clustering records by events allows for trends and patterns to be analyzed quickly and reliably, increasing the potential use of such content for research, advocacy, and accountability. Globally, archiving and preservation of user-generated digital materials documenting human rights abuses and war crimes are increasingly recognized as critical for advocacy, justice, and accountability. For the conflict in Syria, which began in 2011, there are more hours of user-generated content documenting rights violations uploaded to digital platforms than there have been hours in the conflict itself. Whereas some content has been clustered around specific events, such as larger open-source investigations by civil society and documentation efforts, the vast majority of content currently exists as individual unstructured records rather than jointly as clustered events within a relational database. The sheer amount of content and the near constant removals of materials from public channels mean that human rights monitors are in a race against time to preserve content, identify violations, and implicate potential perpetrators. Overcoming challenges related to memory and bias is crucial to this process.