View article

The increasing amount of scholarly literature and the diversity of dissemination channels are challenging several fields and research communities. A continuous interplay between researchers and citizen scientists creates a vast set of possibilities to integrate hybrid, crowd-machine interaction features into crowd science projects for improving knowledge acquisition from large volumes of scientific data. This paper presents SciCrowd, an experimental crowd-powered system under development “from the ground up” to support data-driven research. The system combines automatic data indexing and crowd-based processing of data for detecting topic evolution by fostering a knowledge base of concepts, methods, and results categorized according to the particular needs of each field. We describe the prototype and discuss its main implications as a mixed-initiative approach for leveraging the analysis of …