Collecting, Protecting, and Analyzing Human Rights Data
There will be a discussion of two different examples of liberation technology, one used to collect and protect human rights data, and the other used to analyze it. Martus is a software program that encrypts and remotely backs up data, designed for and used widely by human rights monitors and advocates to protect witness reports and other sensitive human rights data. the focus will be on the security design of Martus and how we addressed the inevitable tradeoffs with usability, as well as the reasons for and consequences of our choice to make Martus free and open source. The legitimacy of human rights advocates is based on their claim to speak truth about a human rights situation, but our ability to know the truth of what is happening on the ground is often severely limited. Founding conclusions on anecdotes or observable events alone can misinterpret both trends over time and the relative distribution of violence with respect to regions, ethnicity or perpetrators. Through the careful application of rigorous statistical methods, the limitations of incomplete data can sometimes be overcome, enabling scientifically-based claims about the total extent and patterns of human rights violations. Also discussed will be how this kind of analysis is enabled by technology, including computational statistical methods, and tools like R and version control to make the analysis auditable and replicable.
Jeff Klingner is a computer scientist with the Human Rights Data Analysis Group at Benetech, where he codes and runs data analysis addressing a variety of human rights questions, including command responsibility of high-level officials in Chad and Guatemala, and mortality estimation in several countries, including India, Sierra Leone, and Guatemala. His technical focus is on data deduplication, machine learning, data visualization, and analysis auditability and replicability. He earned a Ph.D. in Computer Science from Stanford University.