Usher: Enabling Quality Data Collection on Mobile Phones Kuang Chen, Ph.D Student, EECS, UC Berkeley Abstract: Recently, the Director General of the World Health Organization likened the lack of quality health information in low income countries to a “gathering storm”, saying, “to make people count, we first need to be able to count.” All over the world, the pervasiveness of cellular networks is enabling mobile-based data collection. Community health workers armed with commodity phones are starting to visit the most rural and isolated places to promote health and gather data in mobile forms. Among their challenges, data quality ranks high: a recent study found that mobile-based data quality was ten-times worse than dictation to a human operator. In this talk, we will present an overview of Usher, a system to address this challenge. Usher automates form design, filling, and data quality assurance for mobile phones and other non-traditional data entry settings. Rather than relying on domain experts, Usher learns a probabilistic model over form questions using existing form data. Usher then applies this model at every step of the data entry process to ensure high quality. Before entry, it induces a form layout that captures the most important data values of a form instance as quickly as possible. During entry, it dynamically adapts the form to the values being entered, and provides real-time feedback to guide the data enterer toward more likely values, and to avoid errors …