What is the impact of the ‘agency of data’ – defined as the social practices of data collection and analysis in quantitative conflict studies – on practitioners’ perceptions and interpretations of armed conflict? DATAWAR is the first collaborative project to address this research question using a holistic, mixed-methods approach - combining Sciences Po’s expertise in the sociological analysis of International Relations, Ecole polytechnique’s expertise in the use of quantitative approaches to understand social phenomena, and Sciences Po Lille’s expertise in bridging the worlds of academic research, journalism, NGOs, and military officials. Practitioners involved in the coverage, analysis, and management of armed conflict rely more and more on the scientific output from quantitative scholars of conflict to analyse and predict change in contemporary conflict. Furthermore, NGOs and international organizations integrate findings from quantitative scholarship to develop indicators for conflict warning systems. However, there has been few little systematic research regarding the underlying scientific practices of the collection, coding, mathematical analysis and publication of quantitative conflict data – and even less scholarship on the quantitative and qualitative impact of these practices on practitioners’ interpretations of armed conflict and resulting normative conclusions. We argue that these practices tend to exclude dynamics that are more difficult to measure quantitatively, and thus do not reflect the theoretical evolution in the study of conflict of the recent years. Thus, quantitative studies of armed conflict may lead to biased perceptions of war, and perhaps even contribute to linear, deterministic assumptions about the causes of violence among practitioners. To address this lack of research, DATAWAR will analyse the full ‘lifecycle’ of quantitative data on armed conflict. Using a combination of interviews and corpus-based qualitative content analysis, the project will investigate practices of data collection, coding, scholarly analysis, and the discursive transformation of scientific output by government and military officials, NGOs, and journalists. The project will be structured into three main work packages: WP1 (led by Thomas Lindemann) will focus on the ‘production’ of knowledge based on quantitative conflict studies. Contents of the two most renowned scientific journals in quantitative conflict analysis will be sampled and analysed. This analysis will inform interviews with database managers and authors of the most influential research articles to find out about ‘hidden’, informal practices in scientific production. WP2 (led by Frédéric Ramel) analysing the dissemination and reception of this academe knowledge in the media and among private and public actors. In a first step, a corpus of media content will be constructed in order to find out how results from quantitative conflict analysis are communicated in the media, and to what extent these results participate in the dissemination of specific analytical frames and normative agendas. These results will inform research interviews with practitioners (officials of government agencies and international organizations, NGO analysts, private risk advisors, journalists) to find about how they perceive and actively use output from quantitative conflict analysis, including established or new conflict databases. WP3 (led by Eric Sangar) will organize the links with practitioners by collecting their needs and inputs, and assuming responsibility for the entire dissemination process, focusing on the continuous involvement of practitioners throughout the project duration. Beyond ‘traditional’ academic output, this dissemination strategy includes the creation of a web-based interactive tool, the Conflict Database Compass, and a course offering guidance for a more aware use of conflict data by officials, journalists, and NGO staff.