The spread of democracy has been crucial to developing a world order that has facilitated productive economic, social, and cultural growth, yet by almost any measure, democracy is in crisis. I have made a leading contribution to comparative approaches to the study of democratic innovation. I have been to the fore in drawing insights from democratic theory and empirical social science, connecting researchers with practitioners of democracy in government and society. With this fellowship, I will develop interventions to avert the crisis of democracy. Despite continued support worldwide for democracy as a regime, democracy as a practice is suffering. Issues include declining trust in government and political parties, distorted digital communications, and rising populism and polarisation in politics. In a positive response, governments, businesses, and charities are already reimagining democracy. They have designed inventive democratic services and devices that can help sustain democratic order. Some examples include participatory budgeting, randomly selected juries, and different forms of referendums. These social innovations are often supported by civic technologies, open data applications, citizen science, and behavioural nudges such as information cues that increase civic volunteering. Yet we know little about what works beyond case studies. I have taken a leading role in the development of a comparative research agenda. A number of projects such as participedia.net (of which I am a co-investigator and executive member), have begun to collect systematic data on how these devices improve democracy (or do not). Despite the abundance of information, research has yet to take advantage of the analytic potential of data science and new technologies. In the project, I will bring together traditional survey data, and new forms of crowdsourced and real-time data to understand what interventions actually help to sustain rather than hinder democracy. In the first step, the project will push the frontiers of knowledge about what has worked before and what has not. I will use set-theoretic methods, at the cutting edge of comparative analysis in the social sciences to determine the conditions in the past that have been necessary and sufficient for increases in positive democratic behaviours. The method can establish which combinations of conditions in different contexts achieved democratic improvements such as inclusion, learning, deliberation, and support for institutions. At the University of Southampton, I am uniquely positioned among prominent social and computer scientists to lead a multi-disciplinary research team in developing indicators and data analytics for democratic innovation. I will harness available data to provide the necessary information on developing political contexts to guide policymakers in the development and choice of instruments for democratic decision-making. My work will reduce wasted resources in public consultation. These indicators will include new measures of the extent to which debates are consolidating in the public sphere using social media data, argument mapping and opinion polling. Economic indicators of government capacity, as well as indicators of civil society capacity and the levels demand for inputs from citizens will be incorporated to complement those data. The ultimate aim of the project is to use advances in traditional and new forms of data analysis, to work in accordance with the best that democratic theory and political philosophy has to offer. The project will involve agile design of indicator dashboards and complementary social interventions. In conjunction with international and national experts in public engagement, we will deliver field experiments to test feasibility of designs. The fellowship will allow me to lead a multi-disciplinary research agenda developing data science that responds to and integrates the lessons of democratic theory and empirical social science.