We are pleased to announce the publication of Human Rights Violations Reporting Dataset in Data in Brief (Elsevier).
The article, co-authored by Constantinos Djouvas, Nikandros Ioannides, Iosif Kovras, and Christos Christodoulou, introduces a new dataset designed to support research on human rights violations, political violence, and organizational reporting practices.
The Human Rights Violations Reporting Dataset is the first paragraph-level corpus of human rights reports enhanced through advanced text processing and natural language processing techniques. The dataset contains 832,220 paragraphs drawn from reports produced by Amnesty International, Human Rights Watch, the United States Department of State, and the United Nations Working Group on Enforced or Involuntary Disappearances, covering the period 1999–2023.
In addition to the original texts, the dataset includes a wide range of metadata and computational annotations, including named entities, sentiment scores, content classifications, and indicators of different forms of violence and human rights violations. These features enable researchers to conduct large-scale analyses of how violations are documented across countries, time periods, and reporting organizations.
The dataset has been developed to facilitate research in computational social science, political science, human rights, and natural language processing. By making these data openly available, the project seeks to support new forms of interdisciplinary research on human rights documentation, political violence, and enforced disappearances.
The publication forms part of DISACT’s broader commitment to advancing innovative research on human rights violations and creating accessible resources for the scholarly community.
Find the article here:
Human Rights Violations Reporting Dataset (Data in Brief)
Access the dataset here:
Human Rights Violations Reporting Dataset (Zenodo)