Civic Space
Data & Forecasts
Detecting early warnings of civic repression through large-scale text analysis
Funding for the Machine Learning for Peace (MLP) was terminated in February 2025 following the Trump Administration’s cancellation of the Department of Defense Minerva Research Initiative and dismantling of the United States Agency for International Development.
Our final set of civic space data and forecasting updates covers events through November 2024.
Around the world, the fundamental freedoms that allow citizens to organize, speak, and shape politics are increasingly under pressure. Yet most monitoring efforts focus on major crackdowns or headline-grabbing events, missing the day-to-day contestation that truly defines civic space—and often signals deeper political shifts.
The Civic Space dataset fills this gap by using advanced AI and human-supervised methods to track 20 types of civic space events across 65 developing countries. Built on the Machine Learning for Peace (MLP) platform, this dataset draws on an original corpus of more than 120 million news articles from nearly 350 domestic outlets and 30 international and regional sources to measure monthly changes in protests, restrictions, advocacy efforts, legal challenges, media pressures, and other forms of civic activity that shape democratic life.
Explore the Civic Space dataset through this interactive dashboard – examine the civic space event types, track changes over time and across countries.
For more information on the previous use and applications of the data, visit the research section.
