CAFÉ (Collective Assessment & Feedback Engine)


While polling and survey instruments can offer governmental agencies and private industries crucial insights into the effectiveness of public services, policy decisions, and reform initiatives, such methods tend to limit citizen participation and lack innovative tools for generating ongoing and productive public discourse. The Collective Assessment and Feedback Engine (CAFÉ) addresses these shortcomings through the development of an online platform that provides users immediate visual feedback about their position on key issues in relation to other participants while also fostering open-ended dialogue and ongoing community interactions that facilitate more nuanced and generative assessments and understandings of complex issues.

Demonstrating the DIL Approach

Envisioned as a virtual coffeehouse, CAFÉ provides a grassroots, user-friendly, and self-organized environment that empowers the wisdom of crowds to provide the necessary feedback data for institutions and organizations to make informed decisions while also increasing citizen engagement and agency in the decision-making process through accessible and informative visualizations of their participation.

As a project of the CITRIS Connected Communities Initiative, the CAFÉ team is building collaborations with local government agencies, public health researchers, and technological innovators operating in developing regions to improve assessment of interventions in humanitarian assistance, disaster relief, solar energy projects, democracy and governance, and global health.


With support from DIL, CAFE was able to implement three applications of CAFE: in Uganda to evaluate effectiveness of family planning programs, in Mexico to crowdsource policy issues for the June 2015 midterm elections, and in California to solicit feedback on the state’s performance on timely policy issues. These implementations directly influenced changes to the family planning education program in Uganda, identified priority policy issues raised by the National Electoral Institute in Mexico, and revealed disaster preparedness as a priority for Californians which prompted the development of The team has developed a library of statistical models and algorithms for collaborative filtering that streamline and structure feedback on complex issues in the field. Furthermore, the team has also developed a bilingual translation feature (see, a visual- and voice-based feature for low-literacy populations (see, and a feature that allows simultaneous data collection from visual- and text-based systems (e.g., smartphones) and voice-based systems (e.g., feature phones) (see

Lead Researchers

Learn More