The Zooniverse

The remarkable success of Galaxy Zoo as a citizen science project for galaxy classification within a terascale astronomy data collection has led to the development of a broader collaboration, known as the Zooniverse. Activities will include astronomy, lunar science, solar science, and digital humanities. Some features of our program include development of a unified framework for citizen science projects, development of a common set of user-based research tools, engagement of the machine learning community to apply machine learning algorithms on the rich training data provided by citizen scientists, and extension across multiple research disciplines. The Zooniverse collaboration is just getting started, but already we are implementing a scientifically deep follow-on to Galaxy Zoo. This project, tentatively named Galaxy Merger Zoo, will engage users in running numerical simulations, whose input parameter space is voluminous and therefore demands a clever solution, such as allowing the citizen scientists to select their own sets of parameters, which then trigger new simulations of colliding galaxies. The user interface design has many of the engaging features that retain users, including rapid feedback, visually appealing graphics, and the sense of playing a competitive game for the benefit of science. We will discuss these topics. In addition, we will also describe applications of Citizen Science that are being considered for the petascale science project LSST (Large Synoptic Survey Telescope). LSST will produce a scientific data system that consists of a massive image archive (nearly 100 petabytes) and a similarly massive scientific parameter database (20-40 petabytes). Applications of Citizen Science for such an enormous data collection will enable greater scientific return in at least two ways. First, citizen scientists work with real data and perform authentic research tasks of value to the advancement of the science, providing "human computation" capabilities and resources to review, annotate, and explore aspects of the data that are too overwhelming for the science team. Second, citizen scientists' inputs (in the form of rich training data and class labels) can be used to improve the classifiers that the project team uses to classify and prioritize new events detected in the petascale data stream. This talk will review these topics and provide an update on the Zooniverse project.